MetroViz: Visual Analysis of Public Transportation Data

From Cmsc734_f13
Jump to: navigation, search

Contents

TEAM

Team Members

  • Joshua Brulé: jtcbrule@gmail.com
  • Fan Du: fan@cs.umd.edu
  • Peter Enns: peter@umiacs.umd.edu
  • Varun Manjunatha: varunm@cs.umd.edu
  • Yoav Segev: segev@cs.umd.edu

Advisors

  • Michael VanDaniker: mvandani@umd.edu
  • Ben Shneiderman: ben@cs.umd.edu

IMPORTANT DATES

  • October 3, 2013: Draft Proposal
  • October 10, 2013: Revised Proposal
  • November 17, 2013: First Version Demo
  • November 20, 2013: Usability Testing Plan
  • November 21, 2013: Usability Testing Rehearsal and Pilot Study
  • November 22, 2013: First Round Usability Testing
  • December 2, 2013: Second Version Demo
  • December 4, 2013: Second Round Usability Testing
  • December 6, 2013: Draft Paper, Video and Demo
  • December 13, 2013: Final Version of Paper, Video and Demo. Presentation

DATA

  • Blacksburg Transit Data (2010-2013): Bus Stop, Bus Trip, Vehicle ID, Fare Count, Fare Type, Schedule Arrival/Departure Time, Actual Arrival/Departure Time.

REQUIREMENTS

Our primary user, Michael VanDaniker, has expressed his interests in 1) how well bus trips adhere to their schedule at each stop and 2) how many people are affected in a non-adherence event. Our goal is to provide an effective system to visualize the bus transit data of Blacksburg City, and help the transit experts (including Michael VanDaniker) to detect abnormal bus stops and bus trips. In the future, our system could possibly be deployed in the Department of Transportation and assists in schedule optimization.

Mandatory

At a minimum, our visualization will incorporate this data and answer these questions:

  • Which stops make passengers wait for long time?
    • Which stops? - Search for stops
    • How long? - Compare schedule time with actual time
    • How often? - Count
    • How many passengers? - Count by FareCount
  • Potential Reasons
    • Affected by previous stops’ delays? - Search pre stops
    • Due to bad schedule of trips? - Calculate the frequency
    • Due to bus drivers’ personal reasons? - Filter out by WorkName (?)

Our visualization could be divided into three levels:

  • City level: show an overview of all the stops in a city, and support the sort-by-feature functionality to help detecting abnormal stops.
  • Stop level: show the details of a specific stop by time (year, month, week, day), and support comparing functionality among multiple stops.
  • Trip level: show the details of each trip (involving stops along the route), schedule time, actual time, delta, etc.

Desired Features

In the second round of development, we want to provide solution to these queries:

  • Geographical distribution of non-adherence stops? - #Need Geo information#
  • Is an non-adherence caused by the traffic peaks? - #Need traffic info by location and by time#
  • Is an non-adherence caused by emergent situations on the bus or traffic accident? - #Need event info#
  • Is an non-adherence caused by events like holidays? - #Integrate Google to find out#

DEMO

Screen Mockups

Mockup Slides

System Development

  • Visualization: D3.js (browser-based)
  • Data Processing and Server-side: Python/Ruby
  • Database: PostgreSQL/MySQL

Implementation Results

Video, Source Code on Github

MetroViz overview.png

USER STUDY

Target Users

  • Researchers at the Center For Advanced Transportation Technology (CATT), University of Maryland
  • Department of Transportation Experts

Schedules

Round 1

  • Date: Nov. 22, 2013
  • Location: CATT Lab
  • Participants: 3 CATT Lab researchers

Round 2

  • Date: Dec. 6, 2013
  • Location: CATT Lab (off campus lab)
  • Participants: 3 CATT Lab researchers

Subjects

Subjects for our usability testing were researchers from the Center for Advanced Transportation Technology (CATT), University of Maryland. Subjects were recruited by our mentor Michael VanDaniker and volunteered to participate in our usability testing. Subjects had not been involved in our project, and never met us before the usability testing.

Materials

Survey, Task List and Questionnaire

MEETING MINUTES

  • 2013/10/21 | CATT Lab | Joshua Brulé, Fan Du, Peter Enns, Varun Manjunatha, Michael VanDaniker
    • Problems Specifications
      • Which stops make passengers wait for long time?
        • Which stops? - Search for stops
        • How long? - Compare schedule time with actual time
        • How often? - Count
        • How many passengers? - Count by FareCount
        • Geographical distribution? - #Need Geo information#
      • Potential Reasons
        • Affected by previous stops’ delays? - Search pre stops
        • Affected by the traffic peaks? - #Need traffic info by location and by time#
        • Due to bad schedule of trips? - Calculate the frequency
        • Due to bus drivers’ personal reasons? - Filter out by WorkName (?)
        • Due to emergent situations on the bus or traffic accident? - #Need event info#
        • Due to events like holidays? - Use Google to find out
    • Visualization Dimensions
      • Stops (aggregate by RouteName, WorkName, Block, etc.)
      • Schedule time, Actual time, Delta (sort by delta)
      • Population (sort by population)
      • Fare types (sort by fare types)
  • 2013/11/4 | CATT Lab, Joshua Brulé, Fan Du, Peter Enns, Varun Manjunatha, Yoav Segev, Michael VanDaniker
    • Mockup discussion/review
      • "Overview" glyphs: should be configurable (by ridership or adherence); avoid using circles or colors at the overview-level because these are difficult to compare to each other. Michael suggests bar-charts.
      • Consider laying out the overview of all routes vertically (instead of laying out routes in rows, lay out routes in columns, with the map to the left or right)
      • Map: color stops by route (all stops in the same route should get the same color); stops that are part of multiple routes get a `split' symbol with all the colors
      • Individual stop glyphs: Michael suggests replacing the 24-hour clock glyphs with horizontal bar charts; should be configurable by color and height (e.g. user can select color by adherence and height by ridership)
      • Allow for user-defined aggregate glyphs for stops (e.g. see a single bar chart for a single stop aggregated over all of the Mondays for an entire year)
      • "Route view": Since the data is 2 dimensional (trips and stops), bar charts are impractical here; but to compare the different sizes of squares, there should be a grid to help compare different stops on different trips
      • Aggregation: 'average' adherence is useless since very late and very early could average to on-time. Consider sum/average absolute value, sum/average of squares

RELATED WORKS

Analyzing Automatic Vehicle Location (AVL) and Automatic Passenger Count (APC) Data

  • Rancic et al. [1] present a tracking system using AVL data to analyzing city bus transit traffic.
  • Chen [2] introduces a model to simulate bus operation and passenger demand based on AVL and APC data.

Travel Time Predicting

  • Lee et al. [3] introduce a real-time travel time prediction method based on multiple samples of similar historical trajectory.
  • Tiesyte and Jensen [4] propose a method to predict the future movement of a vehicle based on the identification of the most similar historical trajectory.
  • Predic et al. [5] use real-time AVL data and historical data to predict bus motion and bus arrival time.

Route Choosing

  • Nguyen et al. [6] consider buses as moving objects, and use temporal maps to represent the movements of buses in spatio-temporal domain to help passengers choose approatiate routes.
  • Liu et al. [7] propose a bus trip planning system to help passengers choose the most appropriate lines and transfers, based on traffic data.

Trajectory Visualization

  • Tominski et al. [8] use a hybrid 2D/3D display to show the trajectories and associated attributes in their spatial-temporal context.
  • Scheepens et al. [9] improve density maps to help explore trajectories using multiple density fields.

Flows Visualization

  • Guo [10] develops a visualization framework to interactively explore large-scale spatial flows.
  • Cui et al. [11] propose an edge-clustering method to reduce edge crossings and visualize geometry graphs.

Improving Bus Scheduling

  • Kimpel et al. [12] discuss efforts of using the TriMet [13] APC and AVL data to improve buses scheduling.
  • Yu and Yang [14] develop a dynamic holding strategy to optimize the holding strategy.

Schedule Adherence

  • Mai et al. [15] extend the Marey graph by adding schedule adherence and passenger load information to measure transit performance.

Level-of-Service (LOS) Estimation

  • Camus et al. [16] propose a way to estimation the level-of-service (LOS) based on AVL data.
  • Hammerle et al. [17] analyze the AVL and APC data of Chicago Transit Authority to estimate service reliability.

Maps on Mobile Device

  • Wang and Chi [18] introduce a focus+context method to visualize metro map on small displaying area of mobile devices.

Public Transit Services

  • Wmata [19]: Washington Metropolitan Area Transit Authority.
  • TriMet [13]: Public Transit in the Portland Area.
  • BT [20]: Blacksburg Transit provides bus transportation primarily to and from the campus of Virginia Tech.

REFERENCES

  • [1] D. Rancic, B. Predic, and V. Mihajlovic, "Online and post-processing of AVL data in public bus transportation system," WSEAS Trans. Inf. Sci. Appl., vol. 5, no. 3, pp. 229–236, Mar. 2008.
  • [2] W.-Y. Chen and Z.-Y. Chen, "A Simulation Model for Transit Service Unreliability Prevention Based on AVL-APC Data," in Measuring Technology and Mechatronics Automation, 2009. ICMTMA ’09. International Conference on, 2009, vol. 2, pp. 184–188.
  • [3] W.-C. Lee, W. Si, L.-J. Chen, and M. C. Chen, "HTTP: a new framework for bus travel time prediction based on historical trajectories," in Proceedings of the 20th International Conference on Advances in Geographic Information Systems, 2012, pp. 279–288.
  • [4] D. Tiesyte and C. S. Jensen, "Similarity-based prediction of travel times for vehicles traveling on known routes," in Proceedings of the 16th ACM SIGSPATIAL international conference on Advances in geographic information systems - GIS ’08, 2008, p. 1.
  • [5] B. Predic, D. Stojanovic, S. Djordjevic-Kajan, A. Milosavljevic, and D. Rancic, "Prediction of bus motion and continuous query processing for traveler information services," pp. 234–249, Sep. 2007.
  • [6] H. T. Nguyen, C. K. T. Duong, T. T. Bui, and P. V. Tran, "Visualization of spatio-temporal data of bus trips," in 2012 International Conference on Control, Automation and Information Sciences (ICCAIS), 2012, pp. 392–397.
  • [7] J. Liu, Y. Yuan, F. Li, and W. Ding, "Bus Trip Planning Service Based on Real Time Data," in SRII Global Conference (SRII), 2011 Annual, 2011, pp. 530–539.
  • [8] C. Tominski, H. Schumann, G. Andrienko, and N. Andrienko, "Stacking-Based Visualization of Trajectory Attribute Data," IEEE Trans. Vis. Comput. Graph., vol. 18, no. 12, pp. 2565–2574, Dec. 2012.
  • [9] R. Scheepens, N. Willems, H. de Wetering, G. Andrienko, N. Andrienko, and J. J. van Wijk, "Composite Density Maps for Multivariate Trajectories," Vis. Comput. Graph. IEEE Trans., vol. 17, no. 12, pp. 2518–2527, 2011.
  • [10] D. Guo, "Flow Mapping and Multivariate Visualization of Large Spatial Interaction Data," Vis. Comput. Graph. IEEE Trans., vol. 15, no. 6, pp. 1041–1048, 2009.
  • [11] W. Cui, H. Zhou, H. Qu, P. C. Wong, and X. Li, "Geometry-Based Edge Clustering for Graph Visualization," Vis. Comput. Graph. IEEE Trans., vol. 14, no. 6, pp. 1277–1284, 2008.
  • [12] T. J. Kimpel, J. G. Strathman, and S. Callas, "Improving Scheduling Through Performance Monitoring," in in Computer-aided Systems in Public Transport, Springer, 2008, pp. 253–280.
  • [13] "TriMet: Public Transit in the Portland Area." [Online]. Available: http://trimet.org/.
  • [14] B. Yu and Z. Yang, "A dynamic holding strategy in public transit systems with real-time information," Appl. Intell., vol. 31, no. 1, pp. 69–80, Dec. 2007.
  • [15] E. Mai, M. Backman, and R. Hranac, "Visualizing Bus Schedule Adherence and Passenger Load Through Marey Graphs," in 18th ITS World Congress, 2011.
  • [16] R. Camus, G. Longo, and C. Macorini, "Estimation of transit reliability level-of-service based on automatic vehicle location data," Transp. Res. Rec. J. Transp. Res. Board, vol. 1927, no. 1, pp. 277–286, 2005.
  • [17] M. Hammerle, M. Haynes, and S. McNeil, "Use of automatic vehicle location and passenger count data to evaluate bus operations," Transp. Res. Rec. J. Transp. Res. Board, vol. 1903, no. 1, pp. 27–34, 2005.
  • [18] Y.-S. Wang and M.-T. Chi, "Focus+Context Metro Maps," Vis. Comput. Graph. IEEE Trans., vol. 17, no. 12, pp. 2528–2535, 2011.
  • [19] "Wmata: Washington Metropolitan Area Transit Authority." [Online]. Available: http://www.wmata.com/.
  • [20] "BT: Blacksburg Transit provides bus transportation primarily to and from the campus of Virginia Tech." [Online]. Available: http://www.blacksburg.gov/index.aspx?page=791.