Chang, Yung-Ju, Mark W. Newman, Pei-Yao Hung, Manchul Han (2013). Integrating Capture ! Playback into Context-Aware Systems Development. International Symposium of Chinese (ChineseCHI), ACM Conference on Human Factors in Computing Systems (CHI), April 28, Paris, France.
TraceViz: Finding Location Data through Brushing to Support Application Testing
Background: This project was the result of volunteer and course work when I was pursuing my Master's degree at the University of Michigan. I collaborated with a Ph.D. student, Yung-Ju Chang, to design TraceViz for developers of location-based applications. I led the implementation of different filters for developers to refine the search criteria of GPS traces. I also developed a location-based Android app for study participants (e.g., developers) to modify in the evaluation study of TraceViz.
This project was unique for me as we were developing a tool to support developers to create applications. In the design process, I needed to consider not only what developers needed, but also what applications developers might create for app users so that our design could provide proper support.
My role(s): Researcher, Designer, Developer
Collaborator(s): Dr. Yung-Ju Chang, Dr. Mark W. Newman
Method(s):
Skill(s):
Motivation: Location data (e.g., GPS) enable location-based applications (e.g., Yelp) to infer meaningful users' characteristics and behaviors (e.g., interests, favorite places, and jogging routes). Application developers need a way to efficiently find data to test how applications react to different location patterns effectively.
Problem Solving: TraceViz was created to provide a set of filters for finding GPS traces, and a natural 'brushing' interaction for selecting traces. Users could select location traces that pass through a segment of a path (e.g., walking on State street from a coffee shop to a dormitory) by painting the segment using the map interface.
Brainstorm Design
Development
Pilot Testing
Revision
Usability Testing
Outcomes: The lab-based user study (N=8) showed that
- TraceViz provides efficient tools to filter data: participants could use TraceViz to efficiently find location traces needed for testing a sample location-based application.
- The brushing interaction is easy to apply: participants were able to flexibly apply the brushing interaction to narrow down desirable location traces.
The outcomes of this project include multiple scholarly publications (link), an Android application that facilitates social sharing of location data to support group activity (e.g., meeting) as the target of testing, and the TraceViz software that app developers can use to test their location-based apps.
Summary:
Date: 2012 Winter
Research & Design Strategy
(I do have lo-fi sketch...)
Outcomes
Findings & Implications
Systems & Applications