TRCLC 15-7
App-Based Crowd Sourcing of Bicycle and Pedestrian Conflict Data
PIs: Stephen Mattingly, Colleen Casey and Taylor Johnson (University of Texas Arlington)
Summary:
Most agencies and decision-makers rely on crash and crash severity (property damage only, injury or fatality) data to assess transportation safety; however, in the context of public health where perceptions of safety may influence the willingness to adopt active transportation modes (e.g. bicycling and walking), pedestrian-vehicle and other similar conflicts may represent a better performance measure for safety assessment.
Problem:
The rare and random nature of collisions/crashes requires researchers to gather several years of data to produce statistically significant estimates that discard the stochastic variations. Moreover, collision or crash data may be biased and underrepresent actual issues of safety that exist. Finally, problems may only be identified after crashes have occurred.
Research Results:
The entire crowd-source data collection process for conflict analysis included three broad phases. Initially, the study identified key stakeholders related to bicycles and pedestrians and developed a contact list for future involvement in the app snowballing process. Utilizing stakeholders during the various parts of a study can increase the validity of the study by verifying specific needs and present an opportunity to gain additional knowledge from an outside source. Most of them can be characterized into three general groups;
- Those concerned with bicycle and pedestrian safety,
- Those concerned with public/environmental health, and
- Those concerned with city/regional planning and management.
In the next step, the research team developed the functional requirements of the app along with user interface requirements and end user requirements. These led to the app prototype design. The study then tested the prototype and obtained feedback from the stakeholders. The project continued with a beta test and field test of the new app, “Safe Activity.”
In response to the comments received from the field test participants, the team finalized the app with only two user groups. The regular or standard user group can record conflict scenarios, will receive a reminder once a day, and will also receive prompt notification of any conflict recorded their current zip code. This helps the users keep track of hot spot locations around their activity paths. The second user group represents those that will work with the data while at the same time they can also use the app as a regular user. These end users will use the data sharing option of the app to share the required database for any conflict analysis. The database can be shared as a *.CSV file or as a *.KML file, which can be opened in an Excel file or in a Google map file respectively. The admins also have the option to add other users as an admin.
The field test participants recorded a total of 129 conflict records. Out of these conflict records only 9 records happened when a vehicle or a bicyclist has overtaken a bicyclist or a pedestrian respectively. Based on the transportation infrastructure within Arlington, the city lacks shared bicycle lanes around elementary schools. The other 120 conflicts are non-overtaking conflicts. Almost 67% of the overtaking incidents are among bicyclist and vehicles and the rest occur between bicyclists and pedestrians. On the other hand, 83% of the non-overtaking conflicts (120) are pedestrian-vehicle conflicts whereas bicyclist-pedestrian conflicts are at 9% and the remaining 8% are bicyclist-vehicle conflicts.
Results:
The initial field test of the app shows promise with support from many users in continuing to use the app and the app’s effectiveness in mapping conflicts to previously recorded fatalities. Most of the field test users find the app easy to use and the survey questions easy to complete. The user concerns related to locating the conflict location accurately will need to be examined again in the future; however, at this time the researchers lack sufficient data because the research team appears to have failed in its marketing effort to generate sufficient excitement and interest in the app. Marketing efforts will need to be increased in the future to facilitate its broader adoption. After its broader adoption, the data quality and its linkage to crash rates will have to be reexamined.