OIG No. 23-11: Trolley Ridership InspectionJoseph M. Centorino, Inspector General
TO:
FROM:
DATE:
PROJECT:
PERIOD:
Honorable Mayor and Members of the City Commission
Joseph M. Centorino, Inspector General
July 14, 2023
Trolley Ridership Inspection
OIG No. 23-11
October 1, 2021 - March 31, 2023
EXECUTIVE SUMMARY
This inspection was planned and carried out by the City of Miami Beach Office of the Inspector
General (OIG) at the request of Miami Beach Commissioner Ricky Arriola, who expressed
concerns regarding the accuracy of the reported ridership figures for the City's free trolley service.
Those figures, presented to the July 22, 2022 Finance and Economic Resiliency Committee
(FERC) meeting, had been based on an analysis of both manual counts recorded daily by trolley
drivers and the Automatic Passenger Counters (APCs) installed in the trolleys and relied upon by
the City Transportation and Mobility Department in connection with the 2019 Miami Beach Trolley
Passenger Survey data performed by Marlin Engineering, Inc.
The City trolley contractor, Limousines of South Florida, Inc. (LSF), is responsible for operating
the trolley system, and its subcontractor, TSO Mobile by Tracking Solutions Corp. (TSO Mobile),
hired by LSF at the direction of the City, to provide hardware, software, and installation of the
APCs, among other equipment. In addition, the City Transportation and Mobility Department is
responsible for monitoring and determining compliance with the executed Agreement and
Amendments with LSF.
OIG staff researched methodologies to test the reliability of the figures reported to the FERC and
performed an independent inspection to assess the accuracy of the reported ridership data
through visual examination of the limited available footage provided from video cameras installed
in the trolleys. The ridership counts derived from the video footage inspection by OIG staff were
then compared with the reported number of passengers from the APCs and the manual counts
prepared by the trolley drivers. The inspection concluded that the data generated by the APCs
installed in City trolleys and the manual counts prepared by the trolley drivers are unreliable
indicators of the actual ridership for the trolley system, based on the sampled trolleys examined
during March 2023.
Despite multiple requests, the OIG did not receive from the contractor all requested video footage
or daily manual ridership counts prepared by trolley drivers. On some occasions, substitute video
Page 1 of 9
footage of other non-requested trolleys was received instead of the requested footage. The
contractor's failure to provide all the requested video footage created a scope limitation for this
inspection. However, the Transportation and Mobility Department provided an Excel spreadsheet
to the OIG, prepared by TSO Mobile, containing March 2023 trolley drivers' manual counts and
APC counts, so that some additional analysis could be completed by OIG staff.
In the end, the OIG concluded that, due to the insufficiency of the camera footage provided; the
lack of reliable data on which to draw any inference regarding overall ridership; and the skewing
of the data due to the fact that among the nine trolleys tested, the ridership of one trolley was
included three times (which happened to be the trolley which showed the greatest percentage
discrepancy between the reported ridership and that observed by the OIG); no conclusion was
warranted concerning the overall ridership in the system, including any assessment of whether
the ridership count provided to the City was underreported or overreported. The OIG recommends
that the City consider utilizing a more reliable technology for the counting of passengers on its
trolley system.
Written responses to the draft report were received from the City Transportation and Mobility
Department, LSF, and TSO Mobile. Some, though not all, of their comments have been
incorporated into this final report. All responses received are attached to this report in their
entirety.
INTRODUCTION
On May 8, 2014, the City of Miami Beach entered into an Agreement with LSF for turnkey
operation and maintenance services of a Municipal Trolley System in the City of Miami Beach.
According to City Resolution No. 2014-28708, the Agreement provided an option to procure
additional equipment, including, without limitation, Global Positioning Systems with capabilities to
report, Automatic Passenger Counters, Wi-Fi services, and Automated Voice Information
Systems. Pursuant to Article 3, Section 1 of the agreement, the City Administration recommended
that the City secure the optional equipment through LSF, for a sum not to exceed $150,000, during
the initial five-year term of the Agreement, which the Mayor and City Commission approved and
authorized through Amendment No. 1.
Amendment No. 1to the Agreement, approved on September 30, 2014, authorized the execution
of a subcontract between LSF and TSO Mobile to equip the trolleys with the following optional
equipment: Automatic Passenger Counters (APC), Automated Voice Information System, Wi-Fi
services, real-time GPS tracking services (with capabilities to provide mileage, service hours and
ridership reports, and capabilities to provide data in a format that is compatible with Miami-Dade
County's mobile application, "Miami Dade Bus Tracker") and additional automated stop
Page 2 of 9
announcement equipment, with the capability of displaying public advertisements. The OIG
recommends that in the future the City enter into a contract directly with the subcontractor to
enable direct access to the technology provider.
Exhibit B of LSF's contract with TSO Mobile, entered into at the recommendation of the City,
states, "The accuracy of our Automatic Passenger Counting Systems is considered to exceed
that of a single checker on board a bus and comparable to that obtained by two checkers, one
stationed at each door. APC users in the transit systems that have utilized our APC system
consistently express satisfaction with the accuracy of our APC system. In the following, the overall
concurrence value represents the total number of boardings and a lightings' counted by the APC
system compared to the total number of boardings and alightings counted by experienced manual
checkers. The Manual-APC Deviation Range +/- 1 represents the percentage of time the manual
and APC observations were within one (1) of each other. Implementation of our APC system will
result in APC Passenger Count Accuracy levels exceeding 95% concurrence with manual
observations."
Transportation and Mobility Department management emphasized that the comparison of APC
counts to a "single checker on board" is not meant to include the manual trolley driver counts, but
there has been no practice to place single checkers on board other than the drivers that could
provide the basis for a comparison. Therefore, the OIG relied upon its own visual inspection of
riders entering and exiting the trolleys to determine compliance with the Amendment 1 Exhibit B
standard cited above. The OIG's testing posited that, if the APC system report indicates 150
passengers rode a given trolley during the day based on the automatic counters, the manually
observed video ridership counts by OIG staff should range between 149 and 151, most of the
time (95%).
It is understandable that the accuracy of the manual counts performed by the trolley drivers may
be affected by the other responsibilities of the driver throughout the day, including assisting
passengers, driving, communicating with other LSF personnel, etc. Nonetheless, the fact that the
manual counts by the drivers are utilized in place of the APC counts when the latter counts are
unavailable, as indicated below, makes the accuracy of the trolley driver counts fair for
comparison with the OIG counts.
Miami Beach trolley service was suspended on March 26, 2020, following the outbreak of the
COVID-19 pandemic, and eventually resumed at a reduced service level on February 15, 2021.
Therefore, the pre-pandemic analysis involved all trolley operations prior to March 26, 2020, and
post-pandemic represented all trolley operations after February 14, 2021.
It is worth noting that the City does not base its payments to LSF on ridership but rather on
scheduled service hours and established hourly rates as stipulated in Amendment 11 to the
Agreement. Scheduled service hours are defined as the requested number of hours by the City
for services available to the public for transport along designated routes.
Furthermore, the Transportation and Mobility Department stated that fleet size for the trolley
system is based on desired frequency of service and service area, and not on ridership. The
service hours exclude deadhead hours2, interruptions in service, and idle times exceeding 15
minutes. The free Miami Beach trolley covers the following routes: South Beach Loop A and B,
Alighting means to get out a vehicle, especially a train or a bus, according to the Cambridge English Dictionary.
Deadhead hours refers to the measurement of time (in hours) when a vehicle travels from its garage or yard facility to the first
scheduled pick-up point or any time the vehicle travels from the last scheduled drop-off point to its garage or yard facility.
Page 3 of 9
Middle Beach Loop, Collins Express, and North Beach Loop (see Appendix A located at the end
of this report).
The OIG believes that accurate ridership figures would be helpful in the City's overall assessment
of the effectiveness of the trolley program, including the following:
• the number of trolleys that should be operating daily;
• identification of the busiest and least busy routes;
• the most productive hours of operation;
• assessment of the trolley program's appropriate funding level.
WORK PERFORMED
TEST 1
During the July 22, 2022, FERC meeting, the Transportation and Mobility Department presented
post-pandemic trolley service metrics. The following presentation extract indicated that daily
ridership had decreased by 56% from approximately 14,400 pre-pandemic daily riders to 6,300.
MAM.sAH L · RC,GI --ill lift. E, ,
Trolley Service Vetrics
Opr at Hours 18Mk./day 1$ h./day .17%, 1$ h./day
5erce quo.y 5run 30 min 100% 20 min
Rider hip 14,400/5.25 M 6,300/2 3 M ·56% N/A {daily/ annual)
Cost/p0serge $2.30 $3.00 30% N/A
No ol vehicles 25 15$ ·40% 21 n $rvio
Some ridership declines can likely be attributed to the changes in trolley operating hours and
frequency. Prior to the pandemic, trolley service was available 18 hours a day, operating from 6
a.m. to 12 a.m., with 25 vehicles servicing the public at an average interval of 15 minutes.
Following the pandemic, trolley service was initially reduced to 15 hours a day, operating from 8
a.m. to 11 p.m., with 15 vehicles servicing the public at an average interval of 30 minutes. The
Transportation and Mobility Department informed the OIG that trolley service was subsequently
expanded as of October 1, 2022, to include 21 vehicles servicing the public at an average interval
of 20 minutes and operates from 8 a.m. to 11 p.m., which is the current service level.
TSO Mobile representatives informed OIG staff that its automatic counter system produces
ridership reports that can be accessed through its http://app.tsomobile.com website. OIG staff
accessed the corresponding Passenger Counter reports from October 1, 2021, through March
31, 2023, to compute the average daily and monthly ridership. The results indicated an average
Page 4 of 9
of 5,205 daily riders from October 1, 2021, through September 30, 2022, and 6,232 daily riders
from October 1, 2022 through March 31, 2023.
Upon comparing the limited internal TSO Mobile data provided to the OIG with the figures in the
post-pandemic analysis presented by the City Transportation and Mobility Department to the
FERC covering a broader period, the OIG found those figures to be consistent. As the trolley
ridership data before July 1, 2021, was not available, it was not examined by the OIG, so its
accuracy could not be determined.
TEST2
OIG staff initially sought to determine the accuracy of the 14,400 daily ridership figures reported
in the 2019 Passenger Survey conducted by Marlin Engineering Inc. However, TSO Mobile
representatives informed the OIG that the 2019 trolley reports and related video footage were
unavailable, and, therefore, the accuracy of the related reported ridership could not be verified.
Given this limitation, OIG staff had to find an alternative method to assess the accuracy of reported
ridership data and to test the reliability of the passenger counting system.
In response to initial inquiries made by OIG staff, representatives from Transportation America
Inc. (TA), the parent company of LSF, stated that each trolley is outfitted with four interior
cameras. The related video footage is stored on a hard drive within the camera system.
Customers can only enter or exit the trolley through its one door, and the cameras record all its
activities.
In emails to the OIG, TA representatives stated that the video recordings would be stored for
approximately thirty days, thereby limiting the OIG's analysis to 2023 data. However, after
requesting video footage related to the operation of nine trolleys, each operating during a single
day within the prior thirty-day period, TA representatives indicated that video footage was only
available for up to two weeks, although the Exhibit A of Amendment 5 of the agreement states a
minimum period for the storage of recorded data of three weeks.
Consequently, OIG staff accessed the daily Passenger Counter reports from March 13 through
March 19, 2023, within the designated two-week period, and similarly requested video footage of
one day's activities for ten trolleys operating during the same week. The selected sample was
stratified to focus primarily on trolleys with higher reported ridership and to avoid those out of
service during the designated week.
After encountering unexpected difficulties obtaining the requested video footage, TA
representatives provided OIG staff with USB flash drives containing most of the recordings. It is
important to note that one trolley's video footage was unavailable, and substitute video footage
was received for two other trolleys, which resulted in footage for three different days related to a
single trolley, MB08, i.e., the footage provided for three out of the nine trolley days examined
included the same trolley. The OIG utilized all of this footage but must note that, while the repeated
use of footage for one trolley helped to confirm that trolley's inaccurate automatic counter, it would
have been far preferable to have received footage for two other trolleys.
OIG staff examined approximately 135 hours of video footage pertaining to 4,326 riders,
representing 9.4% of the 46,238 riders for the designated week. The number of passengers
entering each trolley during the reviewed day was counted and recorded. The table below
summarizes the identified variances between the TSO Mobile Passenger Counter reports and the
OIG video footage counts:
Page 5 of 9
TSO Mobile Video Variance Variance% Trolley Date Route Report Count
MB08 Thursday, March 16, 2023 South Beach Loop B 297 671 374 125.9%
MB08 Monday, March 13, 2023 Collins Express 301 710 409 135.9%
MB08 Saturday, March 18, 2023 Collins Express 373 643 270 72.4%
MB13 Wednesday, March 15, 2023 Collins Express 557 657 100 18.0%
MB15 Thursday, March 16, 2023 North Beach 417 416 (1) -0.2%
MB18 Sunday, March 19, 2023 Collins Express 659 749 90 13.7%
MB28 Thursday, March 16, 2023 South Beach Loop B 790 388 (402) -50.9%
MB30 Wednesday, March 15, 2023 South Beach Loop B 447 441 (6) -1.3%
MB31 Friday, March 17, 2023 South Beach Loop A 373 311 (62) -16.6%
Trolleys may cover more than one route each day.
** The video count figures include each time the trolley driver entered the vehicle during the examined day as
the APC counts would have done the same. The number of times that the trolley drivers was observed entering
the vehicle during the sampled days above ranged from three for MB13 to 21 for MB30.
The widespread positive and negative variances between the TSO Mobile reports and the OIG
video counts raise concerns about the accuracy of the rider counting process. Each variance,
except for trolley MB15, exceeded the threshold deviation listed in Exhibit B of the TSO Mobile
Contract, some significantly.
In addition, trolleys MB28, MB30, and MB31 were equipped with overhead people sensor APCs3,
which Transportation and Mobility Department staff presented to the OIG as a newer and better
technology than the one used on the other trolleys. As reported in the table above, the associated
variances ranging from 6 riders (-1.3%) to 402 riders (-50.9%), failed to support the assertion that
the newer technology is more accurate than the older technology. It also reported higher ridership
than observed in the video footage related to all three sampled trolleys examined.
Consequently, the OIG recommends that the installation of overhead people sensor APCs for
more trolleys be suspended until the reasons for these deficient counts can be determined and
corrected. If not possible, other more effective counting methods should be researched and
utilized prospectively.
Although the objective of the analysis performed was to assess the reliability of the APCs, the
outcome of which suggests malfunctions leading to inaccurate reporting of ridership figures, the
OIG was also concerned with the inability of TA to provide all requested footage and the
occasional submittal of substitute unrequested footage. For example, video footage for an
additional eight trolleys was requested from TA representatives on April 18, 2023, pertaining to
dates ranging from April 10, 2023 through April 16, 2023, well within the 14 days designated by
TA. On April 27, 2023, the OIG received an email from TA stating that three of the eight requested
videos were obtained, three were unavailable, and they were in the process of obtaining the
remaining two.
The OIG was also informed that trolley drivers perform daily manual ridership counts, which are
compared with TSO Mobile counts, and used when the APC counts are not available. As a result,
the OIG emailed TA representatives on April 27, 2023, requesting the March and April 2023
manual ridership counts, but no response was received.
In the interim, the OIG requested and received an Excel spreadsheet from the Transportation and
Mobility Department, which was previously provided to the City by TSO Mobile, containing the
https ://people-sensing.com/sites/default/files/counter datasheet apc-eco engl.pdf
Page 6 of 9
APC counts and the daily manual ridership counts prepared by trolley drivers operating in Miami
Beach during March 2023. Consequently, the requested April 2023 video footage and daily
manual ridership counts prepared by the trolley drivers were no longer pursued from TA.
The OIG Auditors' examination of the March 2023 Excel spreadsheet found that the APCs for 202
of the 723 line items were blank (27.9%), containing no count whatsoever. This suggests that
APCs installed on many trolleys are not functioning and need repair. After eliminating the 202
blanks from analysis, the OIG compared the remaining 521 reported APC counts with the
corresponding trolley drivers' manual counts. Although the +/- 1 benchmark stated in Exhibit B of
Amendment No. 1 is meant to apply, according to Transportation Department staff, only to a
possible manual account "by a single checker on board a bus" and not to the drivers' manual
counts, it was still concerning to the OIG that only 5 of the 521 line items were within the stated
benchmark.
It should be noted that, on those occasions where the APC devices are not functioning correctly
and do not report any daily riders, the trolley drivers' manual counts are substituted by the
contractor to represent ridership for those days until the devices are recalibrated and repaired, so
it is not illogical to compare the reported APC counts to the trolley drivers' manual counts as a
reasonable measure of their accuracy.
Although not tested by the OIG, it was recommended to the Transportation and Mobility
Department Director in a telephone conversation with the OIG that designated staff monitor the
submitted monthly Excel spreadsheets, if not already done, to ensure that the repair and
recalibration of all malfunctioning APCs occurs within a reasonable time.
Three sampled trolleys, MB28, MB30, and MB31, were equipped with the overhead people sensor
APCs containing cameras which provide a continuous view of the trolley drivers. The
corresponding video footage reviewed by OIG staff found the reliance on manual counts prepared
by trolley drivers questionable, as only one (MB30) of the six observed drivers of the examined
sampled trolleys equipped with overhead people sensor APCs was noticed recording ridership
counts through a clicker hanging from his neck. Although this is not conclusive evidence that
manual counts were not recorded by the other trolley drivers, it does raise questions as to the
origin and accuracy of the manual counts based on trolley driver input.
In a meeting with City staff and TA management, it was stated that the Agreement does not
prescribe how trolley driver manual counts must be completed. Given the number of blanks in the
APC counts for March 2023 (27.9%), the OIG believes that a defined methodology by which the
manual counts are to be performed by the trolley drivers will help result in more accurate ridership
figures reported to the City. As a result, it is recommended that these terms be prospectively
negotiated and added to the next Agreement.
The OIG also compared its video counts from the sampled trolleys examined, each for one day's
operation during the week of March 13 through March 19, 2023, with the manual counts prepared
by the trolley drivers (see the table below). In doing so, it was determined that the trolley drivers'
manual counts were similarly unreliable.
Page 7 of 9
Trolley Date Route Drivers OIGVideo Variance Variance % Manual Counts Count
MB08 Thursday, March 16, 2023 South Beach Loop B 604 671 -67 -11.1%
MB08 Monday, March 13, 2023 Collins Express 758 710 48 6.3%
MB08 Saturday, March 18, 2023 Collins Express 768 643 125 16.3%
MB13 Wednesday, March 15, 2023 Collins Express 985 657 328 33.3%
MB15 Thursday, March 16, 2023 North Beach 466 416 50 10.7%
MB18 Sunday, March 19, 2023 Collins Express 700 749 -49 -7.0%
MB28 Thursday, March 16, 2023 South Beach Loop B 364 388 -24 -6.6%
MB30 Wednesday, March 15, 2023 South Beach Loop B 318 441 -123 -38.7%
MB31 Friday, March 17, 2023 South Beach Loop A 380 311 69 18.2%
In sum, the documentation provided related to March 2023 showed widespread differences
between the APC counts and the trolley drivers' manual counts. Furthermore, both counts
differed, some significantly, from the OIG video counts performed on the daily activities of nine
sampled trolley day-rides. Given the limited documentation received and the small sample in
comparison to the population, the OIG cannot conclude whether the APC counts or the manual
counts prepared by the trolley drivers are more accurate representations of ridership.
The OIG is also concerned that the City may have based some decisions regarding trolley service
on these inaccurate ridership figures. Going forward, the Transportation and Mobility Department
should periodically conduct its own sample ridership inspections with any deficiencies exceeding
a designated percentage promptly addressed with the contractor until being resolved.
Based on the inspection done by the OIG and its communications with the City's Transportation
and Mobility Department, LSF, and TSO Mobile, the OIG recommends that the City consider
utilizing a more reliable technology for the counting of passengers on its trolley system.
cc: Alina T. Hudak, City Manager
Rickelle Williams, Assistant City Manager
Jose R. Gonzalez, Transportation Department Director
Nick Mazorra, COO, Limousines of South Florida, Inc.
Wendy Diaz, Sr. Technical Support Manager, TSO Mobile by Tracking Solutions, Corp.
Rafael Paz, City Attorney
Mark Fishman, First Assistant City Attorney
OFFICE OF THE INSPECTOR GENERAL, City of Miami Beach
1130 Washington Avenue, 6" Floor, Miami Beach, FL 33139
Tel: 305.673.7020 • Hotline: 786.897.1111
Email: CityofMiamiBeachOIG@miamibeachfl.gov
Website: www.mbinspectorgeneral.com
Page 8 of 9
Appendix A
fnr AA[AM/BEACH t TROLLEY
LEGEND
0 NORTH BEACH Loon
0 MlOOU BEAo-4 too
OSOllrHBEACH Loos
_TRANSFER POINTS
%EE o vu
au o #
% 09 %
#s o %±
enw e ts
£
15 1et
UM±
s ~
1$1
st
1t
I!I
PL SOUTH BEACH LOOP - A h co«ii
I
t ls
% i % s
For a full list of trolley stops visit miamibeachtrolley.com
Page 9 of 9
A TRANSPORTATION AMERICA COMPANY
June 21, 2023
Joseph M. Centorino, inspector General Sent Via Email
The Office of Inspector General JosephCcnlormO({[ l11Iamlbeat'hll .go\
City of Miami Beach
1130 Washington Avenue
6th Floor
Miami Beach, FL 33139
Re: LSF Response To Draft OIG Trolley Ridership Inspection Report
Dear Mr. Centorino:
Limousines of South Florida, Inc. (LSF) thanks the Office of Inspector General (OIG) for the
opportunity that was provided to discuss the Draft Trolley Ridership Inspection Report (OIG Report)
during the June 16. 2023 Virtual Meeting, by and between the OIG, City staff, and LSF, and to hereby
submit Responses/Comments in connection thereto.
LSF I S committed to ensuring that the public record provides a factual and accurate account of the
issues discussed in the OIG's Inspection Report, inclusive of perspectives from all relevant parties -the
OlG (as author of the Report), and the City staff and LSF as parties of the Miami Beach Trolley Program
Agreement. To this end, attached please find LSF's Response To Draft OIG Trolley Ridership Inspection
Report.
For purposes of clarity in the public record, LSF's response is provided within the text of the
Draft OIG Report that was provided. Specifically, LSF provides comments in Blue Italics Font in
response to each paragraph of the Ol Draft Report immediately after each paragraph in the
Memorandum, so tbat the OIG can determine which responses will be incorporated into the Final OIG
Report.
Consistent with Section 2-256. Sub-Section (h) of the Code of the City of Miami Beach, LSF
respectfully requests that this complete Response be nonetheless attached to the Final OIG Report so that
LS F's comments that are not incorporated into the OIG Final Report can be available for anyone that
wishes to review LSF's Response/C mments.
If you have any questions, or require any further information, please do not hesitate to advise.
Attachment
2766 NW 62nd Street
Miami, Florida 33147
TeI305.265.3302
Fax30S.26S.3303
1
LSF RESPONSE TD DRAFT DIG TROLLEY RIDERSHIP INSPECTION REPORT
(All LSF Responses/Comments are in Blue Italics Font)
For purposes of clarity In the publIc record. LSF provides comments m Blue ltaiJcs Font In
response to each paragraph of the OIG Draft Report Immediately after each paragraph m the
Memorandum. so that the OIG can determme whIch responses Will be mcorporated mto the Final
O/G Report
ConsIstent Wilt? SectIon 2-256 Sub-Section (h) of the Code of the City of Miami Beach LSF
respectfully fequests lhal this camplele Response be nonetheless attached to the Final OIG
Report so fhat LSF s comments that are not IfIcorporated into the OIG Final Report can be
available fa,. anyone thai vllstles to revIew LSF s Response/Comments
TO: Honorable Mayor and Members of the City Commission
FROM: Joseph M. Centorino, Inspector General
DATE: June 2, 2023 DRAFT
PROJECT: Trolley Ridership Inspection
OIG No. 23-XX
PERIOD: October 1st, 2021 -March 31st, 2023
EXECUTIVE SUMMARY
This Inspection was planned and carried out by the City of Miami Beach Office of the Inspector
General (OIG) at the request of Miami Beach Commissioner Ricky Arriola, who expressed
concerns regarding the accuracy of the ridership figures for the City's free trolley service. Those
figures. presented to the Finance and Economic Resiliency Committee (FERC). had been based
on the Automatic Passenger Counters (APCs) Installed in th e trolleys and relied upon by the City
Transportation and Mobility Department in connection wi th the 201 9 Miami Beach Trolley
Passengel Survey data performed by Marlin Engineering, Inc.
LSF Comments (All LSF Comments are in Blue Italics Font)
Umouslf1es of South FlOrida Inc (LSF) tl1anks the Office of Inspector General (OIG) for the
opportunity thaI was proVided to diSCUSS [he Draft Trolley Rldel shIp Inspection Report (OIG
Repolt) during the June 16. 2023 Virtual Meeting by and between the OIG City staff and LSF.
and to heleby submit Responses/Comments 111 connection thereto LSF IS committed to ensunng
that tt1e publfc leeord plOvldes a factual and accurate dCCOlll71 of the Issues (hscussed III tile OIG s
Inspection Repon. If1C/USlVe of perspectives from All relevant paf11es -the OIG (as author of the
Report) and the City staff and LSF as parties of the MIami Beach Trolley Program Agreement
W,th respect to the substance of Paragraph 1 of the Executive Summary. LSF IS aware of
CommiSSIoner Ano/la CO/1cerns about tl1e accuracy of flclefslup figures fOf tl1e City s flee trolley
serVice In obselVtng Coml1l1ssionel AfI;o/a s specific concerns as expressed from the
CommiSSIon DaIS. LSF notes hiS concerns that the published odershlp figufes may be IIlfiated
In fact. based on filS personal observatIOns. CommISSioner Amola has speculated that ridership
figures were potentially If1 fla ted, and has therefore. quesltoned whether the Cdy should contlllue
investing millIons of dol/a, s mto a Program With lower-than-reported ridership figures
1
2
In tn,s regard, LSF notes thai the OIG s Inspectloll dId not cOl/clucle a,e CIty s trolley nderShll}
f,gures re mflated In (act It was quite the opposite Dunng the designated one-week
observation penoa conducted m March 2023. the DIG found that ndershlp figures were
tll1derreponed by almost sixteen percent (16%) In other words during the desIgnated one-week
penod the Trolley Program transported almost sixteen (16%) more passengers than reported by
the APe eqUipment
LSF respectfully submits that these details are highly probatIve. and should be disclosed
in the opening paragraph of the O/G's Inspection Report.
The City trolley contractor, Limousines of South Florida, Inc. (LSF), is responsible for operating
the trolley system, and its subcontractor, TSe Mobile by Tracking Solutions Corp. (TSe Mobile),
IS the Independent party that Installs and maintains the APCs. In addition, the City Transportation
and Mobility Department IS responsible for monitoring and determining compliance with the
executed Agreement and its Amendments.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF notes thar its sub-contract Agreement with rso Mobile was entered mto at the dIrection of
the City Please I efer to Amendmer1/ No 1 of the Miami Beach Troffey Program Agreement
entered on September 30 2014, which states. /11 relevant part
The Contractor shall provide ItS Optionaf Sel vices by entenng mto a Sub-contract
with TSO Mol)J/e
Accordmgfy. LSF respectfully requests thai the OIG cJanfy thiS paragraph In the Draft OIG Report
as follows
The City lrof/ey contractor Limousines of South Flonda. fnc. (LSF). IS responsible
for operalll1g the troffey system anfi.-( »Its« subcontractor TSO Moblfe by
Trackmg Solullons Corp. (TSO Mobile). IS [he mdependent party »contracted at
the direction of the City« thalll7sla/fs and mamtams the APCs. In addition. the
Citv Transpor1allon and Mobllily Department IS responsible for momtonng and
determmll1g compliance With the executed Agreement and ItS Amendments
(PlmclSA lle,lp II/At ;,-,/ i'fCGlllrlCV LSF.: SlIQC/cst/ow; rnat am stllcken thmugh are rec:nrnmenc7ed to lIP
(/ttlf::/etl lrom Itlto OIG DI<JfI Report ("Jr/rt th(Jse 111m iii e > double arrowed and bold« arE:J suygesled
I y LSF t Il~ dr/r/f'ti mlo the OIG Fmal Report I
elG staff researched methodologies to test the reliability of those figures and performed an
Independent Inspection to assess the accuracy of the reported ridership data through visual
examination of the limited available footage provided from video cameras installed in the trolleys.
The ridership counts derived from the Video footage inspection by elG staff were then compared
with the reported number of passengers from the APCs and the manual co unts prepared by the
trolley drivers. The Inspection concluded that the data generated by the APCs installed in City
trolleys and the manual counts prepared by the trolley drivers are unreliable indicators of the
actual ridership for the trolley system, based on the sampled trolleys examined during March
2023
2
3
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF notes tllat the instal/alion of the APe S IS performed by TSO
To ensure that there IS flO rr1lsunderstandmg as to which entity mstalls the eqUIpment. LSF sllnply
requests this paragraph III the Draft OIG Report be modified as follows.
O/G staff researched methoe/ologles to lest the reliability of those fIgures and
performed an mdependent mspection to assess the accuracy of the reported
fideI shIp data through visual examrnatlon of the Illnited available footage proVided
from Video cameras Installed In the trollevs The ndershlp counts denved from the
Video tootage mspectlon by OIG staff were then compareel with the reported
nllml)er of passengers from the APes »installed by TSO« and the manual
counts prepared by the trolley clnvers The Inspeclion concluded that the data
generated by the APCs mstalled • >by TSO« In Cily trolleys and the manual
coums prepared by the trolley dnvers are unreliable mdicators of the actual
ridership for the trolley system, based on the sampled trolleys examined dunng
March 2023
IP/Pd SP. nOlI" tlldt for U(;CllldC~' L F s . uq(lestlofl~ IIwl alt: -/,,(;)...,.11 rhl()lIgh rife 18c;ommenrieeJ /(1 Ile
le/AI~(' from rhe DIG a'cdr Rt!polt dnc/lho:e Ihat rile .» doub/e arrowed and bold« are slIgge:;red
Vy LSI"" r / 11-' Br/rlea rnf!) //11" OIG F/n<il R8pOr/ )
The OIG did not receive from the contractor, despite multiple requests, all requested video footage
or daily manual ridership counts prepared by trolley drivers. On some occasions substitute video
footage of other non-requested trolleys was received instead of the requested footage. The
contractor's failure to prOVide all the requested video footage created a scope limitation which
raises concerns as to the underlying reasons why it was not furnished to the OIG and what other
defiCiencies may exist. However, the Transportation and Mobility Department prOVided an Excel
spreadsheet, prepared by the contractor, containing March 2023 manual counts and APe counts
to the OIG.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSFapologizes for any delay m the delivery ofrequested mformatlon/data. As noted In the vanoLis
emaIl exchanges LSF generally lesponded to communicaflOns by the OIG either on the same
date 01 withm a reasonable penod Regrettably. the process of obtammg the Video footage was
delayecl eIther because of operational/serVice-related responsIbilities and/or due to
faultv/outdated equIpment.
As LSF notecl elurlng (he Virtual Meetmg With the OIG and City staff. LSF provided the video
footage thac was available and/ol operable as soon as staff was able to retneve It 1SF and TSO
have commul1Icated to the City that the VIdeo eqUipment IS outdated. and shoulcl be replaced As
also dlsctlssed. t/Jere /s new and better passenger cotmtel eqUIpment and video technology
avaIlable and LSF IS wlllmg 10 test these new Soilltlons
Respectfully (he charactenzatlOn that the unavailability of Video footage raises concelf1s as 10
tile LJlldellYll1g reasons wilY If was not furnIsher) to tf7e OIG imd wI/at ot/Jer deficienCies may ex/sf
suggests the pOSSibIlIty of some Improper conduct Tlus 1<; Simply not true nor supportable m the
OfG Repor1 In facl as noted 11"1 Paragf aph 1ofthIS Draft OIG Report. the basis for the InspectIOn
Report is :0 assess the accuracy of the ode/shIp figured of tile City s free trolley service
3
4
specifically whether ndelshlp was Inflated While (he APes and the manual counts have been
determmed unreliable file Drafl OIG Inspection Report did find thai the numbers were
IJncJerteported Thus. there 15 no benefit for LSF to (/elay the proC/UCtIOI1 of requested video
footage. and to suggesl that Any delay raises concerns or that there IS an "/mdellymg reason
fOI such delay IS simply not necess81Y and rmsplaced
As such LSF respectfufly requests that the OIG modify thiS paragraph m the Draft OIG Report as
follows
The OIG did not receive from the contractor, despite multiple requests all
requested video footage 01 dally manual ndership counts prepared by trolley
drivers On some occasions. substitute vIdeo footage of other non-requested
trolleys wa~ received Instead of tile leqllested footage The contractor's failure to
provIde all the requested Video footage created a scope limitation wJ:/i<;R-r-dlSeS
GOHfieF~Y.lAg reasons IfVA.y It waB not furm~GJ..G-.aA.d
whal OIher de.fIGlfJncif!J8 may-exist However. the Transportation and Mobility
Department proVided an Excel spreadsheet, prepared by the contractor
contallling March 2023 manual counts and APe counts to the OIG
I P/~ 15 1/ II-! tf1al f. ' 011 GIl/iiI:, LSF s :;U~JUp.stlOl/:; llill <l1f; ' Illchen tllrouqh ;'lIe {'A(;Ommenr/pti to Ill?
r!eleled r m rnA GIG Dtii{f Rep. If mrl/hose Ihal ill »double arrowed and bold<"< are suqges/erl
IJ LSF to ne ..,(iIIA(/ mIn Int! (JIG Frn.:/! Report )
INTRODUCTION
On May 8, 2014, the City of Miami Beach entered into an Agreement with LSF for turnkey
operation and maintenance services of a Municipal Trolley System in the City of Miami Beach.
According to City Resoluti on No. 2014-28708, the Agreement provided an option to procure
additional equipment, Including, without limitation. Global Positioning Systems with capabilities to
report, Automatic Passenger Counters, Wi-Fi services. and Automated Voice Information
Systems. Pursuant to ArtJcle 3, Section 1 of the agreement, the City Admin istration recommended
that the City secure the optional equipment through LSF, for a sum not to exceed $150,000, during
the initial five-year term of the Agreement. which the Mayor and City Commission approved and
authorized through Amendment No. 1.
LSF Comments (All LSF Comments are in Blue Italics Font)
As noted prevIOusly LSF's SuLl-contract Agreement wrtl1 TSO MobIle was entered Into at the
directIOn of the City Please refer to Amendment No 1 entered on C)eptember 30 ?O 14 whIch
4
5
srates. 1/1 relevant part.
rile COil/racIal shall provIde Its Opl/onal Services by entermg mto a Sub-contract
wltll TSO Mobile '
Accoldmgly LSFrespectfully requests Chal the OIG clarify lhls paragraph in the Draft OIG Report
as follows
On May 8 2014. the City of Miami Beach entered Into an Agreement WIll) LSF for
turnkey Of. erar,on and maintenance services of a MUnicipal Trolley System If1 the
CI/y of Miami Beach According to City Resolullon No 2014-28708 the Agreement
prov/dell an OpllOIl to procur& additIonal eqlllrment at the CIty S leqllesl I17clllcling
without IUHltatlon Global PositIoning Systems wIth capabilitIes to report. AutomatIc
Passenger Counters. Wi-FI selVlces, and Automated VOIce Informatton Systems
Pursuant to Arttcle 3 Section 1 of the agreement the CIty Ad,nlf1lstrallon
ffiGOmmeRflei1 d,rected LSF to that--the City secure the optional equipment through
LSF for a sum not to exceed .$150.000. dUflng the mltial five-yeal term of the
Agreement wluch the Mayor and City CommISSion approved and allthorlzed
through Amendment No 1
IPled (' (ID le fhal tnt aCcurDCI-I... SF S SII9!Wsrl0I1:> IIwl .III s(lIcken Ihrolil/n/r ecommendec! 10 be
ielet/xl 'rom /fl~ DIG Drclft Report wi Ihost.: IhaliJle ··>dollble arrowed and bold« An: suggested
t ly LSf 10 be: dlJlled 1111, rh", OIG Fmal Report I
Amendment No.1 to the Agreement, approved on September 30,2014, authorized the execution
of a subcontract between LSF and TSO Mobile to provide a full turn key Trolley Service , to include
equipping the trolleys with the follOWing optional equipment: Automatic Passenger Counters
(APC). Automated Voice Information System, Wi-Fi services, real-time GPS tracking services
(with capabilities to provide mileage. service hours and ridership reports, and capabilities to
provide data in a format that is compatible with Miami-Dade County's mobile application, "Miami
Dade Bus Tracker") and additional automated stop announcement equipment , with the capability
of displaying public advertisements. The OIG recommends that in the future the City enter into a
contract directly with the subcontractor to help avoid the perception or the possibility that LSF may
exert undue influence on its subcontractor to enhance its ridership fjgures and/or performance
Image.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF does not object to nor partIcularly disagrees With tile OIG ~ suggestion that the CIty should
have direct contractual prrvlfy with an APC prowler to provJcle I idersl?l/> fIgures fOl tI)e City s trolley
program Howevef LSF does oblect to the Inaccurate suggestion that LSF may be exerting
undue Inflllence 011 Us subcontractol to enhance Its I/dersl)!!) "[jUles and/or redormance data '
In III IS legard as plefaced earlier. LSF noces thar the City selected TSO Mobile as Its preferred
APe vendor based all prrce, and per the previously quoted Amendment No 1 to the Miami Beach
Trolley Prog,.am Agreement directed LSF to execute a subcontract With TSO LSF belIeves this
detail I cntlcal to understandrng the orrgll1 of the APC subcontract afl(l tIle nature of LSF's
relationshIP vltl1 T 0
Moreovet LSF specIfIcally asserts that there /s no basis -in perception or In fact -that LSF may
exert Irndue mffuence on the sub-contractor to enhance ndersl7lp numbers '. Indeed, as
5
6
explamed dunng the Virtual Meeting wIth the OrG and CIty staff LSF is compensated based on
~er vice hours -NO T by (lder slllp. ThIS fact IS Dlso actually acknowledged In a subsequent
paragraph In the Drafl OIG Inspectloll Report In fact. dunng the VIrtual Meetmg. City staff also
expressed th,s POll7t and lurtl7el noted that the number of vehicles was/,s not based on ndershlp
figures but ruther on deslled head-way tImes and to mmflnlle bunchmq
It IS therefore frankly. /lomc that the OIG would suggest that LSF would seek to exert undue
Influence andlor enhance ndership figures when there IS clearly no benefit -financIal or otherwIse
-for LSF to do so. In short such language IS unsupportable and unnecessarily inflammatory
LSF therefore respectfllfly requests I/,al the OIG mocllfy thiS paragraph In the Draft OIG Report
as follows.
Amendment No 7 to the Agreement. approved on September 30.2014. authonzed
(hE; execution of a subcontract between LSF and TSO Mobile to proVIde a fuff
wrnkey Trolley SerVIce, to Iflclllde eqwppmg the trolleys wlfh the fol/owm9 optional
eqLllpmenr. Automatic Passenger Coumels (APC), Au/omated VOice Informalion
System W,-F, services real-time GPS trackll7g servIces (with capal)ilitJes to
provide mileage. servIce hours and ndership reports and capabl/llles to proVide
data III a format that IS r:ompatlble with Mlami-D~(1p COLJII/y s mobile application.
Mtcunt Oarie Bus harAel 'J and additional automated stop announcement
eqUIpment. with the capability of displaymg public advertisements. The OIG
recommends> -and LSF concurs -,< that m the future the Cdy enter mto a
Loontl act dlfect/y with the subcontractor », in order to directly oversee and
manage the performance of the deployed techn%gy.« te~1he
{3ef-fiepl-km Of the -pf)ssibility thal--~»a-y-eMm Hnfiue-mfItIeA6&-Gn -#s
suhcontrac/()r10 ef7haAGe-its ndershlfJ H€lLlre-s--and/e( reriefH18ftGe image
(P1erJM' nOTe ttl"'T fOI rlCr;1II "Y LSF , SIIf:}C1eSIIOIiS tilal fire ,101 ~e/lIi1lOuqh 'lIe let ommpneie(i to be
,tculteci .'1 om 1/1f' OIG Draft Reporr ana I/JO~e Inat aIL »double arrowed and bold« 8re sugge~ler:t
b)-LSF to b . £1(1(/':(/11110 Ihl OIG Fin,iI Report)
Exhibit B of the TSO Mobile Contract states, "The accuracy of our Automatic Passenger Counting
Systems IS considered to exceed that of a single cheCKer on board a bus and compara ble to that
obtained by two checkers, one stationed at each door. APC users in the transit systems that have
utilized our APC system consistently express satisfaction with the accuracy of our APC system.
In the following , the overall concurrence value represents the total number of board ings and
al ightlngs 1 counted by the APC system compared to the tota l number of boardings and alightings
counted by experienced manual checkers. The Manual-APC Deviation Ra nge +/-1 represents
the percentage of time the manual an d APC observations were within one (1) of each other.
Implementation of our APC system will result in APC Passenger Count Accuracy levels exceeding
95% concurrence with manual observations_"
LSF Comments (All LSF Comments are In Blue Italics Font)
LSF concurs WIth th,s paragraph
In other words, If the APC system report indicates 150 passengers rode a given trolley during the
day, the manual counts will range between 149 and 151, most of the time (95%). Furthermore,
Alighting means to get out a vehicle, especially a train or a bus, according to the Cambridge English Dictionary.
6
7
both the manual counts and the APC system report counts should be consistently within the same
+/-1 deviation range with the video footage counts performed by OIG staff.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF concurs wIth th,s paragraph
Miami Beach tro lley service was suspended on March 26, 2020. fo llowing the outbre ak of the
COVIO-1g pandemic, and eventually resumed at a reduced service level on February 15, 2021 .
Therefore. pre-pandemic analysiS from LSF involved all trolley operations prior to March 26, 2020,
and post-pa ndemic represented all trolley operations after February 14. 2021.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF concurs with thiS paragraph
It is worth noting that the City does not directly base its payments to LSF on ridership but rather
on scheduled service hours as stipulated in Amendment 11 to the Agreement. Scheduled service
hours are defined as the requested number of hours by the City for services available to the public
for transport along designated routes. The service hours exclude deadhead hours2, interruptions
In service, and idle times exceeding 15 minutes. The free Miam i Beach trolley covers the following
routes: South Beach Loop A and B, Middle Beach Loop, Collins Express, and North Beach Loop
(see Appendix A located at the end of this report).
Among other benefits, accurate ridership figures are Importa nt to help the City Administration to
determine the follOWing:
• the number of trolleys that should be operating daily;
• Identification of the bUSiest and least busy routes through which decisions could be made
to add or reduce the number of trolleys or the hours of operation;
• whether program costs may be offset by gra nt fu nding based on the achievement of
specific ridership targets; and
• assessment of the tro lley program's overall effectiveness and its appropriate funding level.
The first two listed above could impact scheduled service hours, and, therefore, payments
due by the City to LSF.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF generallv concurs wllh the first several sentences of thiS paragraph but agam stresses that
routes ale determined by the City and have nothmg to do with ridership figures. As LSF noted
earliel. and City staff confirmed dUling the Virtual Meetmg With the OIG and LSF, the number of
trolfevs necessary to provide the service was also determll7ed by the City based on troJ/ey route
alignment. established headwayslfrequency of service. and to minimize bunching Thus. the
number of trolleys and service hours are not dependent on Ihe number of passengers usmg the
service and IIdershlp figures have no impact on the TIOI/ey Program sse/vice hours 01 the
amount of payments to LSF
Deadhead hours refers to the measurement of time (in hours) when a vehicle travels from its garage or yard facility to the first
scheduled pick-up point or any time the vehicle travels from the last scheduled drop-off pOint to its garage or yard facility.
7
8
Thel etore LSF respectful/} equests Ihal the OIG make the following modificatIons to thIs
paragraph III (he Draft OIG Report 10 ensure accuracy In the public record as follows
It IS wonh nOlll7g Ihal the City does 1I0t Li/feclly base Its payments to LSF on
ncJershlp bur rathel on scl7eduled servIce hours as stipulated If? Amendment 11 to
tl1e Agreement Scheduled service hours are defme(/ as the requested numbet of
hour s by the CIty fa, services available to the public for transport along designated
routes The servIce hours exclude deadhead hours Ifllerruptlons in selVlce and
Idle Iunes exceeding 15 mmutes »Routes are determined by the City, and the
number of trolleys necessary to provide the servIce was a/so determined by
the City based on established headways and to mmimlze bunching. Thus,
the number of trolleys and service hours are not dependent on the number
of passengers using the service« The free Miami Beach trolley covers the
followmg routes South Beach Loop A and B. Middle Beach Loop Co/lms Express
and North Beach Loop (see Appendix A located a1 the end of thiS report)
Among other benefits. accurate nclershlp figures are Important to help the C,ly
Admlfllstralion to determll1e the followlllg
-- the ntJRIDeI-of lroll~&thal-MJGl-Jld I:le--Gpel:atiflfj-cJail.y-.
• idenllf,catloN-of t-he-busieSl-aoo Je.a.sl OtiS)' f-Gtl/es lhrou§R-wRIGh-deGisiGRB-GOUld '*' made to afitI~ reduce the nvmheF-GI-trolle-ys-or the hours of.epei'8/1tJA-i
• w/7ether proglam costs may be offset by 9rant fundll7g based on the achievement
of specific ndershlp cargets: and
• assessment of the trollev program's overall effecClveness and Its appropnate
funding level
The first two listed---abov~uld impaGt-sGhe!iuled S-€FViGe--h6!l1'S, and.
ther&for~ayments-due-by-lhe City to-kSF.
(PIWlSf' Ilorp thai inl ;J(:(:III']1 V ~SF <; sligpesrloll:l ti1cH dlr~ srllt,Aelll/uOu(/h die rer.ommerllieclln be
(j'.jlelecl ;rorn tf1 DIG Drdfr Report n(f IhoSt! thar file »double arrowed and boJd« <lie s(lgypste[iI, L FIn 'Je dllclerl 1(110 Ihl! il , FIIldl Rp/Jc,!/ )
WORK PERFORMED
TEST 1
During the July 22, 2022 FERC meeting, the City Transportation and Mobility Department
presented post-pandemic trolley service metrics. The following presentation extract indicated that
daily ridership had decreased by 56% from approximately 14.400 pre-pandemic daily riders to
6,300.
LSF Comments (All LSF Comments are ;n B/ue Italics Font)
LSF concurs With tillS pa,aglapl1.
Deadhead hours refers to the measurement of time (in hours) when a vehicie travels trom its garage or yard facility to the first
scheduled pick-up point or any time the vehicle travels from the last scheduled drop-off point to its garage or yard facility.
8
9
Trolley Service
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF concurs with tills Chart
Some ridership declines can likely be attributed to the changes in trolley operating hours and
frequency. Prior to the pa ndemic. trolley service was available 18 hours a day, operating from 6
a.m. to 12 a.m., with 25 vehicles servicing the publ ic at an average interval of 15 minutes.
Following the pandemic. tro lley service was reduced to 15 hours a day, operating from 8 a.m . to
11 p.m ., with 15 vehicles servicing the public at an average interval of 30 minutes. wh ich is the
current service level.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF concurs with tl1ls paragraph
TSO Mobile representatives informed OIG staff that its automatic counter system produces
ndershlp reports that can be accessed th ro ugh its http://app.tsomobile.com website. OIG staff
accessed the corresponding Passenger Counter reports from October 15\ 2021, th rough March
3P\ 2023, to compute the average da ily and month ly ri dership. The results indicated an average
of 5,205 dally riders from October 1'1 2021, through September 30tt , 2022. and 6.232 dally riders
from October 1'! 2022 through March 31 st. 2023.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF (IDeS 1I0t (/Is(lute the OIG S l epresentatlOI1S 111 this paragraph.
Upon comparing the lim ited internal TSO Mobile data provided to the OI G to the fig ures in the
post~pandemic analysIs presented by the City Transportation and Mobility Departm ent to the
FERC covenng a broader penod. the OIG found those figures to be consistent. As the trolley
ridership data before July 1. 2021. was not available, it was not examined by the OIG, so its
accuracy could not be determined.
9
10
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF concurs with (/1/5 paragraph
TEST 2
GIG staff Initially sought to determine the accuracy of the 14,400 daily ridership figures reported
In the 2019 Passenger Survey conducted by Marlin Engineering Inc. However, TSO Mobile
representatives informed the OIG that the 2019 trolley reports and related video footage were
una vailable, and, therefore, the accuracy of th e related reported ridership could not be verified.
Given this limitation, OIG staff had to find an alternative method to assess the accuracy of reported
ridership data and to test the reliability of the passenger counting system.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF (/oes nor (jlspute rhe OIG s replesenlatlons Ifl thiS palagraph.
LSF also adds that as discussed with the OIG and City staff therp. IS new and better passenger
counter eqUIpment and Video technology available and LSF IS wllfmg to test these new solutions
In response to initia l inquiries made by OIG staff, re presentatives from Transportation America
Inc. (TA), the parent company of LSF, stated that each trolley is outfitted with four interior
cameras. The related video footage is stored on a hard drive within the camera system.
Customers can only enter or exit the trolley through its one door, and the cameras reco rd all its
activities.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF Loncurs wllh thiS paragrapl1
In emalls to the OIG, TA representatives stated that the video recordings would be stored for
approximately thirty days, thereby limiting the OIG's analysis to 2023 data. However. after
requesting Video footage related to the operation of nine trolleys , each operati ng during a Single
day within the prior thirty-day period , TA representatives Indicated that Video footage was only
available for up to two weeks.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF generally concurs With thiS paragraph. alth(lugh to ullderseol e a prevIous pomt because the
equipment IS outdated LSF has found fllat the maXInWI7l storage capacity of the memory cal d
was unreliable (0 fOI eease. As such LSF proVIded what was available
LSF tilerefore respectfull) requests that the OIG make (he followmg modifications to the Draft
OIG Report to tfNs paragraph as follows
In emalis to the OIG TA representatives stated Inat the video recoldlllgs would be
srorecl for [JpplO IInatel.'; tlurty (Iavs thereby Illnitlf1g 1/1e OIG s analYSIS to 2023
data Howeve, after lequestmg Video footage related to the operation of nme
(rolleys each operatll1g dunng a s/Ilgle day wilhlll tile pnol tilifty-day penod TA
representatives mdlcated that Video footage was only available for Lip to two
10
11
weeks », as that is the maximum storage capacity of the outdated memory
card.<'<
(P/"fJS/~ nat 111C)II)I.1(; ;11f'1( v LSF sliyges/tolls lIIat I' :IIICkel Ihrouql1 11"" re ommencfeu to h~
IIcl(>larl rrom til OIG Orcllt Report cmd those IhClI .11 U ..;.double arrowed and bold < are suqg sleef
b.. L SF /n [If .l</c/eel ,11/ ,ttlf' )/..., l=,nHI Re{J(Jr/ )
Consequently, OIG staff accessed the daily Passenger Counter reports from March 13 through
March 19,2023, within the designated two-week period , and similarly requested video footage of
one day's activities for ten trolleys operating during the same week. The selected sample was
stratified to focus primarily on trolleys with higher reported ri dership and to avoid those out of
service during the designated week.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF noes not cilspute !lie OIG representations af the met/wc/ology employed In the analysIs.
After encountenng unexpected difficulties obtaining the requested video footage. TA
representatives provided DIG staff with USB flash drives containing most of the recordings. It is
Important to note that one trolley's Video footage was unavailable, and substitute video footage
was received for two other tro lleys, which resulted in footage for three different days rel ated to a
Single trolley, MB08, Le., the footage provIded for three out of the nine trolley days examined
included the same trolley_The OIG utilized all of this footage but must note that, wh ile the repeated
use of footage for one trolley helped to confirm that troll ey's inaccurate automatic counter, It would
have been far preferable to have received footage for two other trolleys.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF does not dIspute Ule OIG s representatIons of the methodology employed in the analYSIS
and agrees lIlat It wou/ci have been far preferable to have received footage for two ather trolleys
OIG staff examined approximately 135 hours of video footage pertaining to 4,326 riders,
representing 9.4% of the 46 ,238 riders for the designated week. The number of pa ssengers
entering each trolley during the reviewed day was counted and recorded. The table below
summarizes the Identified variances between th e TSO Mobile Passenger Counter reports and the
OIG video footage counts:
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF (Ioes 1101 (lispute the OlG 's representallons of the methodoloqy emploved 111 the analYSIS.
Trolley Date Route •
T50
Mobile
Report
Video
Count Va riance Variance %
MBOS Th ursd ay, March 16, 2023 South Beach Loop B 297 I 661 364 I 122.6%
MBDS Monday, March 13, 2023 Collins Express 301 703 402 133.6%
MBOS Saturday, March 18, 2023 Collins Express 373 637 264 70.S%
M B13 Wednesday, March 15,2023 Coll ins Express 55 7 654 97 17.4%
MB15 Thursd ay, Ma rch 16, 2023 North Beach 417 397 ·20 -4.S%
M B18 Sun day, M arch 19, 2023 Collins Express 659 736 77 11.7%
MB28 Thursday, March 16, 2023 South Beach Loop B 790 378 -412 -52.2%
11
12
Wednesday, March 15, 2023 South Beach Loop B 447 -6.0%
Fri day, March 17, 2023 South Beach Loop A 373 -20.4%
Trolleys may cover more than one route each day.
LSF Comments (All LSF Comments are In Blue Italics Font)
LSF doe' flot cltspute tile DIG S repfeSf:l1tatlOlIs contained in the Chart prepared tn connection
wIth the OIG Draft Inspection Report
The widespread positive and negative variances between the TSO Mobile reports and the OIG
performed video counts ra ise concerns about the accuracy of the ri der counting process. Each
variance exceeded the threshold deviation listed in Exhibit B of the TSO Mobile Contract, some
significantly.
LSF Comments (All LSF Comments are in Blue Italics Font)
LS~ generally onCLJrs WIth thIS paragraph. but agam notes for the public record that there was
110 eVIdence of over-reponmg of ndershlp fIgures -whIch IS Ihe pnmary slated reason for the
OIG s Inspectloll Report
LSF thelefore respectfull'f requests (hat the OIG modify thiS paragraph in l/1e Draft O/G Repon
as fOllows
T he WIdespread positive and negative variances between the TSO Mobile reports
and the OIG performed VIdeo counts raise concerns about Ihe accuracy ofIhe oder
COUl1tll1g process Each variance exceeded {he threshold devlaflon fisted In Exhlblf
§ of tile TSO MobIle Contract, some sIgnificantly >However the review did
not support a finding that LSF was over-reporting the number of
passengers <,
,Pit nsp 1]('1(' /11,]1 r.x "ccUlaq L. F S SIIY9csiruns (fIn! I II. '/lICkC'rl I/J(('uqf, arc IOGomrnenrfe I/o be
leIAI!"!(/ f(,'m In l.IG D'df! Report lfld IhOSfJ rhar ,lIe »double arrowed and bold <: are suggested
y LSF II") Of' ir/(/pd I'll th OIG F'ndl RepO/I I
Also concerning is that Trolleys MB28, MB30, and MB31 were equipped with overhead people
sensor APCs'\ which Transportation and Mobility Department staff expected to produce more
accurate results. The use of the overhead people sensor APCs was presented to the OIG as a
newer and better technology than the one used on the other trolleys. As reported in the ta ble
above, the associated variances, ranging from -6 .0% to -52.2%, failed to support this assertion.
Consequently, the DIG recommends that the installation of new overhead people sen sor APCs
for more trolleys be suspended until the reasons for these deficient counts can be determined
and corrected. If not possible, other more effective cou nting methods should be re searched and
utilized prospectively.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF concurs with thIS paragraph and agam adds thaI LSF has commuf)/cated to the OIG and
City staff tiJal Ihere IS new and belter passenger counler equipment and video technology
<: vailable and LSF IS wlflmg /0 test these new solulions
hllps J!people·sensmg._ol11!sites/default!files!counter datasheet apc-eco engl pdf
12
13
Although the objective of th e performed analysis was to assess the reliability of the APCs, the
outcome of which suggests malfunctions leading to Inaccurate reporting of ridership figures, the
GIG was also concerned with the inability of TA to provide all requested footage and the
occasional submittal of substitute unrequested footage. For example, video footage for an
additional eight trolleys was requested from TA representatives on April 18, 2023, pertaining to
dates ranging from April 10,2023 through April 16. 2023, well within the 14 days designated by
TA. On April 27. 2023, the OIG received an email from TA stating that three of the eight requested
videos were obtained, three were unavailable, and they were in the process of obtaining the
remaining two.
LSF Comments (All LSF Comments are in Blue Italics Font)
As noted 117 response 10 othel paragraphs by the OIG dlscussmg Ihe unavailability of some video
footage LSF a93m apologIzes for any delay In the de/lvery of requested mformation/daca As
stated In the variOLls email exchanges. LSF generally responded to communications by the OIG
el1/1el on the ame elate or wlthm a reasonable peflod Regrettably the plOcess of oblammg the
VideO footage was delayed either for operattonal reasons or due to faulty eqUipment As LSF also
/loced dUring the VII/ual Meelmg with the O/G and City stafr LSF provided tile video footage that
was available and/of ope/aNe as soon as staff was able to rerneve it LSF and TSO have
commul7lcated co the CIt} that the vIc/eo eqUipment IS dated and should IJe replaced As also
dlsclIssed the IS new and better passenger COUlJtel eqUipment and Video technology available.
and LSF IS wlllmg to test these new solwlOns.
The GIG was also informed th at trolley drivers perform daily manual ridership counts, which are
compared with TSO Mobile counts . and used when the APC counts are not available. As a result,
the GIG emalled TA representatives on April 27 , 2023, requesting th e March and April 2023
manual ridership counts, but no response was received.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF IIlcorporales and re-states its Comments If7 response to the puor paragraph In the Draft OIG
Report to thiS paragraph as well
In the interim, the OIG requested and received an Excel spreadsheet from the Transportation and
Mobility Department. which was previously proVIded to the City by the contractor, containing the
APC counts and the manual ridership counts for trolleys operating in Miami Beach during March
2023. Consequently, the requested April 2023 Video footage and the March through April 2023
daily manual ridership counts were no longer pursued from TA.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF tncorporates and re-states Its Comments In response to (/7e prior paragraph in the Draft OIG
Report to t/71S paragraph as well
The OIG compared the reported APC counts to the manual counts to determine whether each
figure would be within +/-1 of the other 95 times out of 100 (95% confidence level), as claimed in
Exhibit B of the TSO Mobile contract. The GIG Auditors' examination of the proVided March 2023
Excel spreadsheet found that the APCs for 202 of the 723 line items were blank (27.9%),
containing no count whatsoever. This analysis suggests that APCs installed on many trolleys are
not functioning and need repair.
13
14
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF does nol clispute the OIG s feplesentallons relative to lhe rnetl1odolog{ utrf,zecJ m the
analysIs LSF and TSO have communicated to the CIty that the video equipment IS da/eel. ami
shOUld /Je replaced As also ciisclIssed. there IS new and better passenger counter eqUipment
and video technology avaljable. and LSF IS wIlling to lest these flew solutIons.
The troll eys equipped with the overhead people sensor APCs contained cameras whi ch provide
a continuous view of the trolley drivers. The OIG found the reliance on manual counts
questionable, as none of the three sampled trolleys equipped with overhead people sensor APCs,
showed the dnver recording ridership counts during the days examined. Although this is not
conclusive eVidence that manual counts were not recorded by the trolley drivers, it does raise
questions as to the origin and accuracy of the manual count, especially given the high number of
blanks In the APC counts.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF generally c;oncU/ wtlh thiS paragraph and agall7 adds that LSF has communicated 10 the
OIG and City staffthat there is new and betterpassenger counter eqUIpment and vIdeo technology
avat/able and LSF IS WllIlI1g to test these new solullons
After eliminating the 202 blanks from analysis, the OIG compared the remaining 52 1 reported
APC counts with the corresponding manual counts. It was determined that only 5 of the 521 line
Items compared in the Excel spreadsheet (0.96%) were within th e +/-1 benchmark stated in
Exhibit B of the Agreement. The 0.96% correspondence between the APe and manual counts IS
far less than the stated 95% confidence level listed in the Agreement, fu rther supporting the OIG's
conclUSion that reported ridership counts are unreliable.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF cInes 110t cllspule the OIG S feplesentaffOlls relati.;e to the meil! d%gy utilized III ti,e
analYSIS. and aCids that whIle the results of the mspectlons were unreliable the ridershIp fIgures
vere unde,., eporred LSF and TSO have commul1Icated to the City that the passenger counter
and VIdeo eqUIpment IS outdated and -11ould be replaced As also dlscLissed. there is new and
berte, fJassengel courlTel eqwpment and Video technology availabfe. and LSF is willing to test
thesr:; new solullons
The OIG also compared its video counts from the sampl ed trolleys, each for one day's operation
during the week of March 13 through March 19, 2023, with the manual counts prepared by the
trolley drivers (see the table below). In doing so, it was determined that the manual counts were
similarly unreliable. It is worth noting th at four of the nine sampled trolley days resulted in the
manual counts being lower than the OIG video counts (trolley MB08 on March 16, MB18 on March
19, MB28 on March 16. and MB30 on March 15 .
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF Iflcorporates and re-states Its Comments In response to the pnol paragraph in the Draft ole
Rep0l1 to rhls paragraph as well
14
15
Trolley Date Route •
Drivers
Manual
Counts
OIG
Video
Count
Variance Variance %
MBOS Thursday, March 16, 2023 South Beach Loop B 604 661 -57 -9.4%
I MBOS Monday, March 13, 2023 Collins Express 75 8 703 55 7.3%
MB08 Saturday, March 18, 2023 Collins Express 768 637 131 17.1%
MB13 Wednesday, March 15, 2023 Col lins Express 985 654 331 33.6%
M B1 5 Thursday, March 16, 2023 North Beach 466 397 69 14.8%
M B18 Sunday, March 19, 2023 Collins Express 700 736 -36 -5.1%
MB28 Thursday, March 16, 2023 South Beach Loop B 364 378 -14 -3.8%
M B30 Wednesday, March 15, 2023 South Beach Loop B 318 420 10~ -32.1%
MB31 Friday, March 17, 2023 South Beach Loop A 380 297 83 21.8%
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF do not dispute flu: OIG s representations of file Chart prepared In conneciJon with the
analySIS
The analysis performed In this report mdicates that the City incorrectly relied on the ndership
figures submitted by TA In March 2023. The documentation provided showed widespread
differences between the manual and APC counts, contrary to Exhibit B of the Agreement, and
both differed from the corresponding OIG video counts performed on nine sampled trolley days.
The OIG is also concerned that the City may have based some decisions on these inaccurate
ridership figures, which may also have been reported to other governmental agencies. Going
forward, the Transportation and Mobility Departm ent should periodically conduct its own sample
ndershlp inspections with any defiCienCies exceeding a deSignated percentage promptly
addressed With the contractor unti l being resolved.
LSF Comments (All LSF Comments are in Blue Italics Font)
LSF cannol opme on whether the CIty incorreclfy relied 011 ndershlp data sublmtted by LSF m
Marcn of 2023 As noted preVIously the lmrellability of some data IS genel ally based on oLltdated
eQUipment and technologv In thIs regard, LSF has communicated to tile OIG and City staff that
there IS new and bettel passenger counter eqUIpment and Video technologv available and LSF
IS willmg to test these new solutions.
LSF w1her obsel ves [hat the Agreement between the CIty does not prescnbe any method for the
collectIon 0 manual ridership counts LSF f/.ll1her submits that manLial co Lints provIded by LSF
dnver s durmq the deSIgnated obselvatlon penod were approximately 9 4% which is consistent
WIth I17clustry standards and entIrely reasonable given Ihe other Important duties aSSigned to
trolley dovers In Ih,s regard LSF I elteraces Its concer n wlCh the suggestion thaI the CIty
ornrnJs~ Ion rna !Iave /)aser/ some deCIsions on lI1i1cr;uratp. "ders/llp figIIres ..
Whde not clea,.,y Slated tne GIG suggests that the Cdy CommISSion may have allocated millIons
of ta" dollars to fund Cl progl am With Inflated ndersl1ip flgwes T/7I5 IS Just not accurate as the O/G
Inspec(loli revealed thaI rldel shIp figures dUring Ihe deSignated ()bservallon period were
(/nael'reported not mflated In addllion. as noted above. ndershlp figures have not feaiured
prnmmently If) fhe CIty S cieCJSIOIlS regalcfmg trolley "ervlce IIOUlS elllC! payments to LSF
As to the OIG ~ l.onCI3I11S d(Il1e City tJaSlllQ some {ieclsIOf)S on maCCl/rate flclersi7lp fIgures agam
CIty staff confIrmed dwrnq the Vtrwal Meetrng with tl1e DIG an(1 LSF that the number of vehIcles
15
16
used fOI roules and tne number' of servIce hours requested IS not based whatsoever on fideI ship
figures Rathel CIty scaff has stated that the number of trolleys determined for servIce by the City
IS based on eSlalJl,shed headways and to mmllnize bunchrng Thus. the number of trolleys and
selVICe hours ale not dependent on tl,e number of passengers USing the service
LSF therefore respectfully req(Jests thaI the OIG make the followmg modifications to thiS
palagraph m the OIG Draft Report as follows.
The analySIS-perfermed II~ thiS report IncHGates-tltat RP-GHymcerr-ectlY-Felied-on
1J1e flderSA-"rflglJres StJ{)/-mftM by-TAm MarGR-~; The documentatIon prOVided
SilO wed widespread diffel ences between the manual and APe counts contrary to
E< "hlbll B of the Agreement. and both differed from the cOITespondlng O/G video
counts pel10rmed 011 f7lf1e sampled trolley days T-he O/G-+&-aIs~d-tAat
the City may-ha~&Ei-l:iemo decISIons 011 Ihoso inaccurate rldorshlp figures.
wJ"IlGh maya/Sf) hav eeen-H3-~ to ()tI~1 -gevemmemal-age-nGie& GOing
forward the Transportation .mel Mobility Department should periodically conduct
It ~ own sample nderS!1lp InspectIons With any defiCienCIes exceeding a deSIgnated
pel centage promptly addressed With the contractor untlf resolved
IPI, r 0; 1101 11'I/I,il I ("lIre/l ~ L .F <; SUljQCSIIOI1.<; 111.11 CII !':lflUi('1l /llIOUl/I1.'11 I()( IlmmCIl(/QrlIO 110
1C!IOled r, Hn /ill' OIG 1)-[1111':1]0011 ami 'nose (tml "118 >double arrowed and bold - -are SUClqos!eri
tlV L F,O In Ad Il]n Ill! '1/1 ' O/G F/fla/ Ref)ort I
Joseph M. Centorino, Inspector General Date
Mark D. Coolidge, Chief Auditor Date
cc: Alina T. Hudak, City Manager
Rickelle Williams, ASSistant City Manager
Jose R. Gonzalez, Transportation Department Director
Appendix A
LEGEND -
--O ~·... ~
O s,oun, !lACH lOOp · A , ..
16
1
Alonso, Elisa
From:Wendy Diaz <wdiaz@tsomobile.com>
Sent:Tuesday, July 11, 2023 2:50 PM
To:Centorino, Joseph
Cc:Dan Ocampo; Mark Gilmore; Rick Dunn (GPST)
Subject:Re: TSO Mobile Comment
[ THIS MESSAGE COMES FROM AN EXTERNAL EMAIL ‐ USE CAUTION WHEN REPLYING AND OPENING LINKS OR
ATTACHMENTS ]
Joseph,
I wanted to give you some comments on the current Audit for the City of Miami Beach. I wanted to make sure that it
was understood that our hardware's performance is at times affected by issues the Trolley's may have. As an example
our APC system requires an event of Door Open and Close so that we know they were passengers picked up. If the
system does not receive the door open and close it will ignore the count that was just done. There are quite a few of the
Trolleys that currently have the door mechanisms not working properly, thus the APC is not working properly. TSO is
responsible to provide the service to visualize the results of the data obtained by the equipment, process any hardware
failure/replacement when hardware is still under warranty
(12 months) or offer an upgrade to the latest technology, in order to resolve issues and/or improve results. Any issues
caused by tampering are not covered under warranty and it is the responsibility of LSF to maintain that portion
(mechanic and working of the Trolleys). We have on many occasions helped LSF with these issues, because we value the
City and we are all working together. TSO provides the hardware and installation of the same.
I do know that the City has trolleys still running the older technology for the Passenger Counters. The city has purchased
some replacements that will be the newest technology and should be installed soon. We will install them in Trolley
MB05 and MB07. Once they are installed you can re audit these for accuracy and notice the difference.
Thank you for your time.
Best Regards,
Wendy Diaz
TSO Mobile | a GPS Trackit company
Sr. Technical Support Manager
wdiaz@tsomobile.com | 1‐877‐477‐2922 Ext. 1109
Transportation Department Response In Red Line To Draft Report
Page 1 of 8
Honorable Mayor and Members of the City Commission
Joseph M. Centorino, Inspector General
June 2, 2023 DRAFT
Trolley Ridership Inspection
OIG No. 23-XX
October 1st, 2021 - March 31st, 2023
EXECUTIVE SUMMARY
This inspection was planned and carried out by the City of Miami Beach Office of the Inspector
General (OIG) at the request of Miami Beach Commissioner Ricky Arriola, who expressed
concerns regarding the accuracy of the ridership figures for the City’s free trolley service. Those
figures, presented to the Finance and Economic Resiliency Committee (FERC), had been based
on an analysis of both manual (i.e. driver) counts and the Automatic Passenger Counters (APCs)
installed in the trolleys and relied upon by the City Transportation and Mobility Department in
connection with the 2019 Miami Beach Trolley Passenger Survey data performed by Marlin
Engineering, Inc.
The City trolley contractor, Limousines of South Florida, Inc. (LSF), is responsible for operating
the trolley system, and its subcontractor, TSO Mobile by Tracking Solutions Corp. (TSO Mobile),
is the independent party that installs and maintains the APCs. In addition, the City Transportation
and Mobility Department is responsible for monitoring and determining compliance with the
executed Agreement and its Amendments with LSF.
OIG staff researched methodologies to test the reliability of those figures and performed an
independent inspection to assess the accuracy of the reported ridership data through visual
examination of the limited available footage provided from video cameras installed in the trolleys.
The ridership counts derived from the video footage inspection by OIG staff were then compared
with the reported number of passengers from the APCs and the manual counts prepared by the
trolley drivers. The inspection concluded that the data generated by the APCs installed in City
trolleys and the manual counts prepared by the trolley drivers are unreliable indicators of the
actual ridership for the trolley system, based on the sampled trolleys examined during March
2023. Based on the OIG’s inspection of the limited video footage available, the APCs appear to
be under-reporting actual ridership figures while the driver counts appear to be over-reporting
actual ridership figures.
Commented [WR1]: To OIG: Should reference date when
figures were presented to the FERC
Commented [WR2]: Which figures?
Draft Report
Trolley Ridership Inspection
June 2, 2023
Page 2 of 8
The OIG did not receive from the contractor, despite multiple requests, all requested video footage
or daily manual ridership counts prepared by trolley drivers. On some occasions, substitute video
footage of other non-requested trolleys was received instead of the requested footage. The
contractor’s failure to provide all the requested video footage created a scope limitation which
raises concerns as to the underlying reasons why it was not furnished to the OIG and what other
deficiencies may exist. However, the Transportation and Mobility Department provided an Excel
spreadsheet, prepared by the contractorTSO Mobile, containing March 2023 manual counts and
APC counts to the OIG.
INTRODUCTION
On May 8, 2014, the City of Miami Beach entered into an Agreement with LSF for turnkey
operation and maintenance services of a Municipal Trolley System in the City of Miami Beach.
According to City Resolution No. 2014-28708, the Agreement provided an option to procure
additional equipment, including, without limitation, Global Positioning Systems with capabilities to
report, Automatic Passenger Counters, Wi-Fi services, and Automated Voice Information
Systems. Pursuant to Article 3, Section 1 of the agreement, the City Administration recommended
that the City secure the optional equipment through LSF, for a sum not to exceed $150,000, during
the initial five-year term of the Agreement, which the Mayor and City Commission approved and
authorized through Amendment No. 1.
Amendment No. 1 to the Agreement, approved on September 30, 2014, authorized the execution
of a subcontract between LSF and TSO Mobile to provide a full turnkey Trolley Service, to include
equipping the trolleys with the following optional equipment: Automatic Passenger Counters
(APC), Automated Voice Information System, Wi-Fi services, real-time GPS tracking services
(with capabilities to provide mileage, service hours and ridership reports, and capabilities to
provide data in a format that is compatible with Miami-Dade County’s mobile application, “Miami
Dade Bus Tracker”) and additional automated stop announcement equipment, with the capability
of displaying public advertisements. The OIG recommends that in the future the City enter into a
contract directly with the subcontractor to enable direct access to the technology provider. to help
avoid the perception or the possibility that LSF may exert undue influence on its subcontractor to
enhance its ridership figures and/or performance image.
Exhibit B of the TSO Mobile Contract states, “The accuracy of our Automatic Passenger Counting
Systems is considered to exceed that of a single checker on board a bus and comparable to that
obtained by two checkers, one stationed at each door. APC users in the transit systems that have
utilized our APC system consistently express satisfaction with the accuracy of our APC system.
In the following, the overall concurrence value represents the total number of boardings and
Draft Report
Trolley Ridership Inspection
June 2, 2023
Page 3 of 8
alightings1 counted by the APC system compared to the total number of boardings and alightings
counted by experienced manual checkers. The Manual-APC Deviation Range +/- 1 represents
the percentage of time the manual and APC observations were within one (1) of each other.
Implementation of our APC system will result in APC Passenger Count Accuracy levels exceeding
95% concurrence with manual observations.”
In other words, if the APC system report indicates 150 passengers rode a given trolley during the
day, the manual counts will range between 149 and 151, most of the time (95%). Furthermore,
both the manual counts and the APC system report counts should be consistently within the same
+/- 1 deviation range with the video footage counts performed by OIG staff.
Miami Beach trolley service was suspended on March 26, 2020, following the outbreak of the
COVID-19 pandemic, and eventually resumed at a reduced service level on February 15, 2021.
Therefore, pre-pandemic analysis from LSF involved all trolley operations prior to March 26, 2020,
and post-pandemic represented all trolley operations after February 14, 2021.
It is worth noting that the City does not directly base its payments to LSF on ridership but rather
on scheduled service hours and established hourly operating rates as stipulated in Amendment
11 to the Agreement. Scheduled service hours are defined as the requested number of hours by
the City for services available to the public for transport along designated routes. Furthermore,
fFleet size for the trolley system is based on desired frequency of service and service
areacoverage, and not on ridership. The service hours exclude deadhead hours2, interruptions in
service, and idle times exceeding 15 minutes. The free Miami Beach trolley covers the following
routes: South Beach Loop A and B, Middle Beach Loop, Collins Express, and North Beach Loop
(see Appendix A located at the end of this report).
Among other benefits, Aaccurate ridership figures are important to help the City Administration to
determine the following:
• the number of trolleys that should be operating daily;
• identification of the busiest and least busy routes through which decisions could be made
to add or reduce the number of trolleys or the hours of operation; ; • whether program costs may be offset by grant funding based on the achievement of
specific ridership targets; and
• assessment of the trolley program's overall effectiveness. and its appropriate funding
level.
The first two listed above could impact scheduled service hours, and, therefore, payments
due by the City to LSF.
Ridership figures are not used to determine the number of trolleys that should be
operating, hours of operation, grant eligibility, or the appropriate funding level for the
program.
WORK PERFORMED
1 Alighting means to get out a vehicle, especially a train or a bus, according to the Cambridge English Dictionary.
2 Deadhead hours refers to the measurement of time (in hours) when a vehicle travels from its garage or yard facility to the first
scheduled pick-up point or any time the vehicle travels from the last scheduled drop-off point to its garage or yard facility.
Draft Report
Trolley Ridership Inspection
June 2, 2023
Page 4 of 8
TEST 1
During the July 22, 2022 FERC meeting, the City Transportation and Mobility Department
presented post-pandemic trolley service metrics. The following presentation extract indicated that
daily ridership had decreased by 56% from approximately 14,400 pre-pandemic daily riders to
6,300.
Some ridership declines can likely be attributed to the changes in trolley operating hours and
frequency. Prior to the pandemic, trolley service was available 18 hours a day, operating from 6
a.m. to 12 a.m., with 25 vehicles servicing the public at an average interval of 15 minutes.
Following the pandemic, trolley service was reduced to 15 hours a day, operating from 8 a.m. to
11 p.m., with 2115 vehicles servicing the public at an average interval of 230 minutes, which is
the current service level.
TSO Mobile representatives informed OIG staff that its automatic counter system produces
ridership reports that can be accessed through its http://app.tsomobile.com website. OIG staff
accessed the corresponding Passenger Counter reports from October 1st, 2021, through March
31st, 2023, to compute the average daily and monthly ridership. The results indicated an average
of 5,205 daily riders from October 1st, 2021, through September 30th, 2022, and 6,232 daily riders
from October 1st, 2022 through March 31st, 2023.
Upon comparing the limited internal TSO Mobile data provided to the OIG to the figures in the
post-pandemic analysis presented by the City Transportation and Mobility Department to the
FERC covering a broader period, the OIG found those figures to be consistent. As the trolley
ridership data before July 1, 2021, was not available, it was not examined by the OIG, so its
accuracy could not be determined.
TEST 2
OIG staff initially sought to determine the accuracy of the 14,400 daily ridership figures reported
in the 2019 Passenger Survey conducted by Marlin Engineering Inc. However, TSO Mobile
representatives informed the OIG that the 2019 trolley reports and related video footage were
unavailable, and, therefore, the accuracy of the related reported ridership could not be verified.
Draft Report
Trolley Ridership Inspection
June 2, 2023
Page 5 of 8
Given this limitation, OIG staff had to find an alternative method to assess the accuracy of reported
ridership data and to test the reliability of the passenger counting system.
In response to initial inquiries made by OIG staff, representatives from Transportation America
Inc. (TA), the parent company of LSF, stated that each trolley is outfitted with four interior
cameras. The related video footage is stored on a hard drive within the camera system.
Customers can only enter or exit the trolley through its one door, and the cameras record all its
activities.
In emails to the OIG, TA representatives stated that the video recordings would be stored for
approximately thirty days, thereby limiting the OIG’s analysis to 2023 data. However, after
requesting video footage related to the operation of nine trolleys, each operating during a single
day within the prior thirty-day period, TA representatives indicated that video footage was only
available for up to two weeks.
Consequently, OIG staff accessed the daily Passenger Counter reports from March 13 through
March 19, 2023, within the designated two-week period, and similarly requested video footage of
one day’s activities for ten trolleys operating during the same week. The selected sample was
stratified to focus primarily on trolleys with higher reported ridership and to avoid those out of
service during the designated week.
After encountering unexpected difficulties obtaining the requested video footage, TA
representatives provided OIG staff with USB flash drives containing most of the recordings. It is
important to note that one trolley’s video footage was unavailable, and substitute video footage
was received for two other trolleys, which resulted in footage for three different days related to a
single trolley, MB08, i.e., the footage provided for three out of the nine trolley days examined
included the same trolley. The OIG utilized all of this footage but must note that, while the repeated
use of footage for one trolley helped to confirm that trolley’s inaccurate automatic counter, it would
have been far preferable to have received footage for two other trolleys.
OIG staff examined approximately 135 hours of video footage pertaining to 4,326 riders,
representing 9.4% of the 46,238 riders for the designated week. The number of passengers
entering each trolley during the reviewed day was counted and recorded. The table below
summarizes the identified variances between the TSO Mobile Passenger Counter reports and the
OIG video footage counts:
Trolley Date Route *
TSO
Mobile
Report
Video
Count Variance Variance %
MB08 Thursday, March 16, 2023 South Beach Loop B 297 661 364 122.6%
MB08 Monday, March 13, 2023 Collins Express 301 703 402 133.6%
MB08 Saturday, March 18, 2023 Collins Express 373 637 264 70.8%
MB13 Wednesday, March 15, 2023 Collins Express 557 654 97 17.4%
MB15 Thursday, March 16, 2023 North Beach 417 397 -20 -4.8%
MB18 Sunday, March 19, 2023 Collins Express 659 736 77 11.7%
MB28 Thursday, March 16, 2023 South Beach Loop B 790 378 -412 -52.2%
MB30 Wednesday, March 15, 2023 South Beach Loop B 447 420 -27 -6.0%
MB31 Friday, March 17, 2023 South Beach Loop A 373 297 -76 -20.4%
Total: 4,214 4,883 -669 -15.8%
Formatted Table
Draft Report
Trolley Ridership Inspection
June 2, 2023
Page 6 of 8
* Trolleys may cover more than one route each day.
The widespread positive and negative variances between the TSO Mobile reports and the OIG
performed video counts raise concerns about the accuracy of the rider counting process. Each
variance exceeded the threshold deviation listed in Exhibit B of the TSO Mobile Contract, some
significantly. However, it is important to note that the OIG’s limited review did not support a finding
that APCs wereas over-reporting the number of passengers as the APCs appear to be under-
reporting the number of passengers based on the video footage available.
In addition,Also concerning is that Trolleys MB28, MB30, and MB31 were equipped with overhead
people sensor APCs3., which Transportation and Mobility Department staff expected to produce
more accurate results. The use of the overhead people sensor APCs which was presented to the
OIG as a newer and better technology than the one used on the other trolleys. As reported in the
table above, the associated variances, ranging from -6.0% to -52.2%, failed to support that the
newer technology is more accurate than the older technologythis assertion. Consequently, the
OIG recommends that the installation of new overhead people sensor APCs for more trolleys be
suspended until the reasons for these deficient counts can be determined and corrected. If not
possible, other more effective counting methods should be researched and utilized prospectively.
Although the objective of the performed analysis was to assess the reliability of the APCs, the
outcome of which suggests malfunctions leading to inaccurate reporting of ridership figures, the
OIG was also concerned with the inability of TA to provide all requested footage and the
occasional submittal of substitute unrequested footage. For example, video footage for an
additional eight trolleys was requested from TA representatives on April 18, 2023, pertaining to
dates ranging from April 10, 2023 through April 16, 2023, well within the 14 days designated by
TA. On April 27, 2023, the OIG received an email from TA stating that three of the eight requested
videos were obtained, three were unavailable, and they were in the process of obtaining the
remaining two.
The OIG was also informed that trolley drivers perform daily manual ridership counts, which are
compared with TSO Mobile counts, and used when the APC counts are not available. As a result,
the OIG emailed TA representatives on April 27, 2023, requesting the March and April 2023
manual ridership counts, but no response was received.
In the interim, the OIG requested and received an Excel spreadsheet from the Transportation and
Mobility Department, which was previously provided to the City by the contractor, containing the
APC counts and the manual ridership counts for trolleys operating in Miami Beach during March
2023. Consequently, the requested April 2023 video footage and the March through April 2023
daily manual ridership counts were no longer pursued from TA.
The OIG compared the reported APC counts to the manual counts to determine whether each
figure would be within +/- 1 of the other 95 times out of 100 (95% confidence level), as claimed in
Exhibit B of the TSO Mobile contract. The OIG Auditors’ examination of the provided March 2023
Excel spreadsheet found that the APCs for 202 of the 723 line items were blank (27.9%),
containing no count whatsoever. This analysis suggests that APCs installed on many trolleys are
not functioning and need repair. When TSO identifies that APC devices are not functioning
correctly and in need of repair, TSO uses drivers’ manual counts for those days until the devices
are recalibrated and repaired.
3 https://people-sensing.com/sites/default/files/counter_datasheet_apc-eco_engl.pdf
Draft Report
Trolley Ridership Inspection
June 2, 2023
Page 7 of 8
The trolleys equipped with the overhead people sensor APCs contained cameras which provide
a continuous view of the trolley drivers. The OIG found the reliance on manual counts
questionable., as none of the three sampled trolleys equipped with overhead people sensor
APCs, showed the driver recording ridership counts during the days examined. Note that the
Agreement does not prescribe the manner by which driver manual counts must be completed.
Although this is not conclusive evidence that manual counts were not recorded by the trolley
drivers, it does raise questions as to the origin and accuracy of the manual count., especially
given the high number of blanks in the APC counts.
After eliminating the 202 blanks from analysis, the OIG compared the remaining 521 reported
APC counts with the corresponding manual counts. It was determined that only 5 of the 521 line
items compared in the Excel spreadsheet (0.96%) were within the +/- 1 benchmark. stated in
Exhibit B of the Agreement. The 0.96% correspondence between the APC and manual counts is
far less than the stated 95% confidence level listed in the Agreement, further supporting the OIG’s
conclusion that reported ridership counts are unreliable.
The OIG also compared its video counts from the sampled trolleys, each for one day’s operation
during the week of March 13 through March 19, 2023, with the manual counts prepared by the
trolley drivers (see the table below). In doing so, it was determined that the manual counts were
similarly unreliable. It is worth noting that four of the nine sampled trolley days resulted in the
manual counts being lower than the OIG video counts (trolley MB08 on March 16, MB18 on March
19, MB28 on March 16, and MB30 on March 15).
Trolley Date Route *
Drivers
Manual
Counts
OIG
Video
Count
Variance Variance %
MB08 Thursday, March 16, 2023 South Beach Loop B 604 661 -57 -9.4%
MB08 Monday, March 13, 2023 Collins Express 758 703 55 7.3%
MB08 Saturday, March 18, 2023 Collins Express 768 637 131 17.1%
MB13 Wednesday, March 15, 2023 Collins Express 985 654 331 33.6%
MB15 Thursday, March 16, 2023 North Beach 466 397 69 14.8%
MB18 Sunday, March 19, 2023 Collins Express 700 736 -36 -5.1%
MB28 Thursday, March 16, 2023 South Beach Loop B 364 378 -14 -3.8%
MB30 Wednesday, March 15, 2023 South Beach Loop B 318 420 -102 -32.1%
MB31 Friday, March 17, 2023 South Beach Loop A 380 297 83 21.8%
Total: 5,343 4,883 460 8.61%
The analysis performed in this report indicates that the City incorrectly relied on the ridership
figures submitted by TA in March 2023. The documentation provided showed widespread
differences between the manual and APC counts, contrary to Exhibit B of the Agreement, and
both differed from the corresponding OIG video counts performed on nine sampled trolley days.
The OIG is also concerned that the City may have based some decisions on these inaccurate
ridership figures, which may also have been reported to other governmental agencies. Going
forward, the Transportation and Mobility Department should periodically conduct its own sample
ridership inspections with any deficiencies exceeding a designated percentage promptly
addressed with the contractor until being resolved.
Commented [GJR3]: Not relevant. Accuracy of drivers’
counts are irrespective of APC counts.
Commented [GJR4]: Not sure as to the relevance of this
paragraph or what it is trying to convey as Exhibit B does
not call for APC counts to be compared to driver counts.
Formatted Table
Commented [GJR5]: However, TSO’s monthly exercise of
blending both data sources ultimately yields more accurate
ridership figures as shown by the results of the two
independent OIG tests.
Commented [GJR6]: As noted previously, the trolley
service level since inception has been based on initial policy
decisions regarding desired frequency of service, service
area (i.e. routes), and hours of operation, not on ridership.
Ridership has not been a criteria in current or prior service
levels nor a factor in terms of payments to LSF. Payments
are based solely on number of service hours and
contractually established hourly operating rates minus any
performance penalties including service interruptions and
on-time performance penalties.
Draft Report
Trolley Ridership Inspection
June 2, 2023
Page 8 of 8
___________________________ ___ ________
Joseph M. Centorino, Inspector General Date
___________________________________ ________
Mark D. Coolidge, Chief Auditor Date
cc: Alina T. Hudak, City Manager
Rickelle Williams, Assistant City Manager
Jose R. Gonzalez, Transportation Department Director
Appendix A