Traffic Incident Heat Maps

From Cmsc734_08
Jump to: navigation, search

quick link to Meeting minutes

Project report

upload here

Current idea

Traffic Management Centers log hundreds of traffic incidents each day, so gathering an understanding of the data can be a daunting task. The usual process of determining high accident locations begins with a request from a citizen who has already noticed a number of accidents in a particular location. The traffic engineers then query all of the accidents associated with that location and determine a course of action. This is a rather passive approach; conditions must be bad enough that citizens complain to the agency.

We will develop a tool that supports a more proactive approach to hot spot detection. The tool will generate heat maps for a given region based on a number of filters. The filtering mechanism will help users not only identify where accidents happen but under what conditions they occur. For example, when compared to the entire data set, there may be relatively few accidents at a given location, but when you consider only the morning rush hour during rainy conditions, that location may suddenly become much more of a concern.

This tool will work with live data as well, allowing traffic managers to better coordinate their units in the field to deal with the region as a whole.

Project Members

Basic interface (mobile)

Basic interface is intended for the general public. It will feature bigger and fewer controls, give access to simple and well-understood discovery tools (many people understand graphs, but fewer understand treemaps). The main goal is to make application scalable so that it can be used on a mobile phone. More on the interface here.

Members:

  • alt Andreea Olea Andreea Olea
  • alt Ken Knudsen Ken Knudsen
  • alt jhj at cs Joonghoon Lee

Complex interface (for traffic engineers)

Complex interface will be used by traffic engineers or other traffic management professionals. It will have a slightly different interface than the mobile version since it will offer more controls for dynamic querying. The users will have access to more complex visualizations (spiral graph), will be able to save the queries for later (as bookmarks). Since the traffic management may use some sensitive data, the access to the interface will be restricted (username/password screen). More details here.

Members:

  • alt Darya Filippova Darya Filippova
  • alt Michael VanDaniker Michael VanDaniker
  • alt kristw at cs Krist Wongsuphasawat

The teams will continue meeting together to discuss progress and further direction. The teams will each produce a technical report of a publishable quality at the end of the semester.

Meeting minutes

Meeting minutes

Contact

Michael Pack (CATT Lab)

Email: PackML at umd.edu

Office: 301-403-4628

Links

Center for Advanced Transportation Technology (CATT Lab) data and code sponsor

Maryland Strategic Highway Safety Plan 2006-2010 includes a part about developing tools that will help solving the road safety problems

MDSHA (MD State Highway Administration)

User interfaces for highway traffic management (HCI Lab) traffic-related projects and publications from the previous years

CHART (Coordinated Highway Action Response Team) interactive map, road closures, speed data for MD

Flex tutorial by Michael VanDaniker - a "Hello, world" type of tutorial

Flex Quick Starts - a nice "How To" for Flex

References

  1. "Temporal, Geographical and Categorical Aggregations Viewed through Coordinated Displays" by Anna Fredrikson, et. al. (also here). New Paradigms in Information Visualization and Manipulation. 1999. 26-34.
    • The paper explores various aggregations on the incident data in the state of Maryland.
  2. "Interactive visual tools to explore spatio-temporal variation" by Andrienko, N., Andrienko, G. Proceedings of the working conference on Advanced visual interfaces. 2004.
    • The paper presents CommonGIS - a tool that helps to analyze spatio-temporal data. The authors outline a set of questions that can be solved using maps and demonstrate solutions in CommonGIS.
  3. "Interactive data visualization using focusing and linking" by Buja, A., McDonald, J., Michalak, J., Stuetzle, W. Proceedings of the 2nd conference on Visualization '91. 1991.
    • The paper establishes the importance of focusing and linking. In spatial displays, selecting a point on the map should highlight the items in detailed view - the principle we have adapted in our tool.
  4. "Vehicle Speed Information Displays for Public Websites A Survey of User Preferences" by Bhamidipati, P., Plaisant, C. August 2000.
    • The study described in this paper helped us identify color schemes that would be most beneficial for users.
  5. "Survey of websites providing real time traffic information on maps" by Plaisant, C. March 2000.
    • Catherine's review of existing traffic information systems helped us identify a set of best practices for developing interfaces in this domain.
  6. "A Structure of Problem-Solving Methods for Real-Time Decision Support in Traffic Control" Molina, Martin / Hernandez, Josefa / Cuena, Jose International Journal of Human-Computer Studies 1998 v.49 n.4 p.577-600
    • The paper presents first a domain-independent model for real-time decision support as a structured collection of problem solving methods. Then, it is described how this general model is used to develop an operational version for the domain of traffic management.
  7. "Interactive Visualization of Serial Periodic Data Visualization" / Carlis, John V. / Konstan, Joseph A. Proceedings of the ACM Symposium on User Interface Software and Technology 1998 p.29-38
    • This paper presents a spiral visualization technique, which display data along a spiral to highlight serial attributes along the spiral axis and periodic ones along the radii. It shows interesting way of using spiral in 2D and 3D.
  8. "Visualizing time-series on spirals". Weber, M, Alexa, M, Muller, W. Information Visualization, 2001. INFOVIS 2001. IEEE Symposium on 22-23 October 2001 Page(s):7 - 13
    • We have implemented the spiral diagram described in this paper to show how incidents build up over an arbitrary time range.
  9. "A Rank-by-Feature Framework for Interactive Exploration of Multidimensional Data". Seo, J and Shneiderman,B. Information Visualization, 4, 2 (June 2005), 99-113. (HCIL-2004-31)
    • We have adopted the rank by feature framework to help users identify interesting locations on the map.
  10. "Exploratory Data Analysis With Categorical Variables: An Improved Rank-by-Feature Framework and a Case Study" Jinwook Seo and Heather Gordish-Dressman. International Journal of Human-Computer Interaction, Volume 23, Issue 3 December 2007 , pages 287 - 314
    • Since our dataset contains many categorical variables, we need some kind of ranking criteria for categorical variables. Ranking criteria from the previous paper (HCE) may be meaningful for quantitative variables, but not appropriate for categorical variables. This paper is a continued work of HCE which presents a ranking criteria to evaluate relationships between variables when at least one variable is categorical using chi-square test.
  11. "Finding communication hot spots of location-based postings" Lemmelä, Saija-Maaria / Korhonen, Hannu J. Proceedings of ACM CHI 2007 Conference on Human Factors in Computing Systems 2007 v.2 p.2549-2554
    • The works presented in this paper show use of a semi-transparent heat map over a map view to visualize posting density on the area of interest.
  12. "How to make heat maps" by Corunet. August 6, 2006.
    • This blog post describes how Corunet, a Spanish web development company, renders heat maps. Their product displays a heat map on top of a web page showing where visitors click. We have adapted the algorithm described in this paper to work in our domain.

Others

Pictures

Here's a sketch of what the heat map visualization may look like:

The heat map sketch