Introduction, Task Flow,
Work Distribution

Prototype

Method & Test Measures

Results

Discussion

Appendix

Problems we Encountered in User Testing:

Screen 1: Front Page

• Map or no map? Our first iteration had a real-time traffic map on the front page. This confused users, leading them to infer that our system only calculated real-time traffic, rather than predictive traffic patterns. One user said she'd rather not have a map at all until she specifies a trip, so we ditched the front-page map altogether for the third test.
• Time format? One participant wanted a more specific time format to enter in her departure/arrival time. No format was specified on the prototype.

Screen 2: Predicted traffic page

• Confusion with predicted traffic data: Two participants did not even understand that this map showed predicted traffic, based on historical data. One participant said "I don't believe that as of Thursday, you can tell me what will happen on Saturday... you just have no idea what will happen in the future." We realized from this comment that we needed to label more specifically the historical, predictive nature of the data being mapped.
• Not understanding that data is predicted on a 5-minute granularity: The one participant who did understand that our data was based on historical estimates assumed that it's one big average rather than an average for every 5-minute period over the course of a week. She said she wouldn't trust the system for travel planning, because the travel time estimate would be identical to Mapquest's. Again, there is clearly a need for more descriptive labelling.
• Distrust of "predicted traffic": One participant said she would never trust a predicted traffic estimate, because accidents and anomalies seem common. All three participants said they would more frequently use the real-time view.
• Not noticing the slider: Two of the three participants ignored the time slider completely, not even realizing it was there or had any purpose. The one participant who noticed it only used the slider briefly, to get a quick glance at the change in traffic if she altered her departure time. This problem may have resulted from the logistics of a paper prototype, where a slider is less noticeable than when a mouse is at the user's disposal. Nonetheless, the slider needs to be labelled or otherwise made more prominent.
• Usefulness of "Save as Commute" button: Not one of our participants clicked on this button. Perhaps this resulted from the construction of our tasks? This needs further testing.
• Ability to view directions for alternate routes: Our first iteration did not allow users to view directions for alternate routes, only the fastest one. Participant B said she'd like to print alternate routes to plan for contingency cases where traffic conditions change dramatically en route.

Screen 3: Real-time traffic page

• Distraction from the traffic alerts: In the first iteration of our real-time traffic page (see photo), we included a series of traffic alerts relevant to the route. This distracted our users, especially since the alerts consumed half a page.
• Not interacting with the map: None of our participants bothered to click on the real-time map, which could have showed them details of traffic accidents. Perhaps this is a limitation of the paper prototype, which is less conducive to mouse interaction than a real system. Still, we might add icons to represent accidents, etc.
  

Screen 4: Directions popup window

• Print and Done: Participant A expressed annoyance at not having a "Print" or "Done" button. We added it for the second and third tests, and it seemed well-utilized by our participants. Both of the other participants used the buttons to complete their tasks.
  

Screen 5: Login screen

• No problems encountered.
  

Screen 6: Registration screen

• Does this save my trip automatically? Participants were confused by what was remembered automatically once they registered. We had overlooked this detail in our design. If a user starts by entering some trip, and then registers, will the trip automatically save to his/her profile? Participants suggested this functionality would be ideal.

Screen 7: Personalized front page

• Cookie-based recognition: One participant expressed annoyance at having to sign in each time. She'd rather have the system save a cookie and immediately recognize her.
• Default to preferred commute: Participants said they'd like the front-page map to immediately show their preferred route, rather than a general Bay Area map.

Screen 8: Profile screen

• Real name: One participant said she'd like to have her real name, not just a screen name, saved in her profile so that the system could greet her accordingly.
• Several saved routes: Participant A, the flexible commuter, suggested she'd like to able to save several routes and immediately look at them for real-time traffic.