Homework 10Solutions/Comments, =55.9, =18.1 / Due: Mon. 22 November 2010
PUBLIC MASS TRANSPORTATION
CE361 HW 10- 1 -
Dear Consultant:
Members of the Board of Directors for the Mythaca Bus Company want to balance the desires of the traveling public, public officials, and taxpayers. Several issues will soon come before the Board, and the Board may call upon (i.e., hire) you to help them make informed decisions.
Please complete the exercises below completely and clearly. Show your calculations clearly. You must submit your assignment as a member of a group of 3-5 CE361 students. You will need that many to collect the data for Problem 1. HW from groups smaller than three will not be accepted. Signatures of all group members must appear on the top page of the work submitted.
1.Testing a Passenger Information System. One year after the previous vendor was replaced, the
new transit passenger information system has been installed and is being tested. A passenger can find the estimated time of departure (ETD) for buses on routes that serve the bus stop at Beering Hall on University Street (Quick Code BUS389) in three ways:- Enter “BUS389” at MyRideWEB at
- Send the text message“BUS389” to “25252”, and getting a response from RouteShout.
- Read the message Board at the bus stop
The MBC Board of Directors has been told that the three information media all use the same source data (from the GPS bus tracking system), but that the source data are processed in different ways and may give inconsistent information to passengers. To provide information useful to MBC, do the following:
(1)Synchronize your watches.
(2)Every two minutes for 10 minutes, acquire the Estimated Time of Departure (ETD) for buses due to serve BUS389 from all three media listed in the bullet list above. Please use “military time”, e.g., 1852 instead of 6:52PM.
(3)Enter the data you collect in the worksheet provided. Each line in the worksheet should apply to the same bus. For example, if the MyRideWeb entries in Columns B-D are for a bus on Route 15 Tower Acres, the entries in Columns F-H and J-L in the same row must be for the same bus. If those data are not available through a given medium, enter “na” into the corresponding ETD cell(s)
(4)Using the “Route Maps & Schedules” feature at gocitybus.com/myrideweb.html, estimate the Scheduled Time of Departure (STD) for each bus due to serve BUS389. Also note the headway for that route. Enter the STD and headway values for each row in the worksheet to which it applies. (There will probably be multiple queries that produce data for a particular bus. This is why each row should pertain to only one bus due to serve BUS389.)
(5)By direct observation, enter the Actual Time of Departure (ATD) for each bus in your database, even if the ATD occurs after your ten minutes of data collection is finished. Enter the ATD value for each row in the worksheet to which it applies. (There will probably be multiple queries that produce data for a particular bus. This is why each row should pertain to only one bus due to serve BUS389.)
- (20 points) Analyze the data that you collected. For example, answer questions such as:
- Is the passenger information consistent between the media? Elaborate.
Will ETD values vary between media much (if at all) for each time of query?
There were 552 useable observations in the worksheets submitted for this HW. From these data, three pairwise differences were computed:- (ETD by MyRideWeb) – (ETD by Message Board)
- (ETD by text) – (ETD by MRW)
- (ETD by text) – (ETD by Message Board)
< -10 / 1
-10
-9 / 1
-8
-7
-6 / 2
-5
-4 / 1
-3 / 3 / 1
-2 / 10 / 7 / 3
-1 / 46 / 19 / 16
0 / 144 / 113 / 75
1 / 96 / 65 / 84
2 / 19 / 15 / 23
3 / 14 / 10 / 9
4 / 5 / 3 / 1
5 / 1 / 2
6 / 2
7 / 0
8 / 3 / 1
9 / 5
10 / 0 / 1
>10 / 2 / 1
Total / 352 / 238 / 214
comparisons seldom differ by more than 2 minutes. If the absolute value between ETDs is used, the average absolute difference for the three pairwise comparisons are 70.4 seconds, 58.7 sec., and 57.2 sec., respectively.
- How accurate are the ETD values with respect to the ATD values? If the ETD values change during your ten minutes of observations, is there any pattern to those changes that you can detect or hypothesize?
Does the ETD tend to get later as queries are repeated at 2-minute intervals?
Are the ETD values distributed around (some earlier than, some later than) the eventual ATD?
The three plots that follow indicate how ATD-ETD varies for each medium as ATD-ToQ (Time of Query) changes. For MRW messages more than 25 minutes before ATD, the ETD value is always the same as, or sooner than, the ATD. The outlier at which ATD-ETD=-27 occurred when the MRW indicated the next bus was 30 minutes away, when it was actually only 3 minutes away. The five outlying points to the upper right were produced during Saturday service.
The plot for the Message Board at BUS389 shows a high degree of accuracy (ATD-ETD=0), even for large values of ATD-ToQ. Again, the ETD is more likely to be earlier than the ATD. Most of the points where ETD is later than the ATD are within 3 minutes of the ATD.
The plot below for the Route Shout text message service also has a lot of points with ATD-ETD=0. The low-lying point at ATD-ToQ=28 was when the bus arrived eight minutes before the text service said it would, but 3 minutes after MRW said it would.
- Does scheduled headway affect the accuracy of the ETD with respect to the ATD?
Plot the data for each headway separately, or side-by-side in the same plot.
Because MRW had the most ATD and ETD observations, the plot below is for ETD as headway varies. Headway = 30 minutes has the widest range, but headways greater and less than 30 do not show any strong trend. Some teams did not seem to know what a “schedule(d) headway” is.
What other questions come to mind? (This would be a good time for your team to brainstorm.) If you use the worksheet to support your analysis by adding columns or charts, provide clear explanations for the MBC Board to consider.
Patterns to look for:
Are ATD values distributed around the STD values?
What is the distribution of ATD-STD values? See summary table below. “na” = data not available
ETD by: / ETD<ATD / ETD=ATD / ETD>ATD / naMyRideWeb / 104 / 111 / 273 / 64
Message Board at BUS389 / 70 / 62 / 237 / 183
RouteShout text / 51 / 48 / 142 / 311
ATD<ToQ / ATD=ToQ / ATD>ToQ / na
9 / 27 / 483 / 33
ATD<STD / ATD=STD / ATD>STD / na
137 / 54 / 322 / 39
Some teams gave good comments distinguishing the three media from user perspective.
Special note: To bus people, the plural of “bus” is “buses”, not “busses”.
- (20 points) Send by email to the worksheet that you have created. Rename the worksheet by giving it the name of the person who is sending it, e.g., smith.xls. Also submit a hard copy of the first page of the worksheet with your report. If the worksheet is longer than one page, do not submit the entire worksheet.
Most worksheets were either done well or could be salvaged with some effort. Surprisingly few teams used their worksheets to support their analyses. Many teams just looked at their data and made some statements. Any bus detected during the ten-minute observation period was supposed to have its ATD value recorded; some teams did not do so. Also, some teams missed an observation time.
- Performance Measures. Money Magazine used percent of the service area pop living within 0.25 mile of a bus route as one of the measures to rank cities.
- (10 points) Is this a measure of effectiveness or of efficiency? Explain.
It is a measure of effectiveness, because it has to do with doing the right thing, not necessarily doing it well. (FTE 539)
- (10 points) Using the 2009 data available for Lafayette CityBus at Indiana Public TransitAnnual Reports, propose a better effectiveness measure for Money Magazine to use. Explain why it is better. (This would be a good time for your team to brainstorm.)
Providing service to an area, but not having the service used, is not the best objective. The “Productivity”, “Performance/Service Effectiveness”, and “Financial Performance” sections for CityBus on page 93 of the 2009 Indiana Public Transit Annual Reports contain a mixture of effectiveness and efficiency measures, if the definitions on FTE 539 are followed. Even under “Performance/Service Effectiveness”, there are efficiency measures. A reasonable separation by categories is:
Effectiveness / EfficiencyTotal Passenger Boardings 4,741,647
Total Vehicle Miles 1,770,275
Revenue Vehicle Miles 1,618,129
Revenue Vehicle Hours 140,045
Passenger Trips per Capita 38.54
Locally Derived Income $4,634,585.14 / Operating Expense per Total Vehicle Mile $5.72
Operating Expense per Passenger Trip $2.13
Passenger Trips per Total Vehicle Mile 2.68
Operating Subsidy Ratio 57%
Locally Derived Income Per Operating Expense $0.46
Fare Recovery Ratio 22%
Operating Subsidy $5,794,467.53does not fit into either category. My favorite effectiveness measure is Passenger Trips per Capita, because it captures service used (passenger trips ) and the potential for service use (population).
- (10 points) Using the 2009 data available for Lafayette CityBus at Indiana Public Transit Annual Reports, propose an efficiency measure for Money Magazine to use. Explain why it is an efficiency (not effectiveness) measure.
I prefer an efficiency measure that includes the amount of service used (i.e., passenger trips) and economic efficiency. Therefore, my choice would be Operating Expense per Passenger Trip $2.13. A measure of output (e.g., ridership) relative to input (e.g., operating expense) makes the most sense.
- (10 points) For the efficiency measure you developed in Part C, calculate its value for the Bloomington IN Public Transportation Corporation. Is Bloomington PTC more efficient than Lafayette CityBus?
For Bloomington PTC, the same efficiency measure (on page 37 of the 2009 Indiana Public Transit Annual Reports is Operating Expense per Passenger Trip $1.87. The CityBus value is $2.13, so Bloomington wins this one.