Mobility Data Contest sponsored by Google

Results

1st Place: Fine-grained Mobility Monitoring Tool

2nd Place: GPS Tracking of Activities in Washington DC

3rd Place: Internet Data Usage as a Predictor of MobiSys Best Paper

Idea

Planning an experiment that relies on mobility-related data, deriving the results and communicating them to a broader public is not trivial. Hence, we want to challenge you to show us your experience in dealing with mobility experiments. This is why, with the generous help from Google, we are organizing the Mobility Data Contest during the 3rd ACM HotPlanet 2011 workshop. The goal is to explore different approaches in mobility data mining - starting from the very beginning, i.e., designing the mobility experiment and finishing with processing the data and sharing your results with a broader audience.

Rules

The contest will be officially launched at the beginning of the HotPlanet workshop (June, 28th). Only people registered for the MobiSys conference and/or associated workshops can participate in the contest. The mobility data has to be gathered and processed during the MobiSys conference (June, 28th - July, 1st). Nevertheless, we let you prepare your mobility experiment ahead, such that you can get ready with your essential equipment, data processing and visualization suite, way before the contest starts. In order to participate in the contest you don’t have to register ahead. It is sufficient if you send us, before midnight June the 30th (EST), your mobility data (URL to the file containing the data will be sufficient) and a 5-slide presentation (in pdf) that includes the following:

• mobility experiment details (slide 1)
• data description/stats (slide 2)
• your data processing and visualization tool chain (slide 3)
• your results (slide 4)
• conclusion/future work (slide 5)

Please note that the 5-slide presentation and the data is a bare minimum. You are more than welcome to submit your animations, videos, etc. This will obviously increase your chances in winning the competition! The more creativity the better!

After the contest your mobility data will be made public in the Crawdad repository. Hence, you must address the privacy issues properly, i.e., all the subjects have to be informed about your experiment, all personal data, MAC, Bluetooth, IP addresses, etc. have to be anonymized. The winners of the contest will be announced at the last day of the MobiSys conference (July, 1st) and all the submitted presentations will be made public on the workshop's website.

Evaluation

The jury will consist of workshop organizers and special guests. The names will be announced at the opening of the contest. The evaluation criteria are the following:

• originality of the mobility experiment
• data cleanness/privacy issues addressed correctly
• statistical significance
• visualization techniques/clarity
• research potential

There will be no rank-ordering of the criteria - each criterion will have the same weight in the final score.

Prize

Thanks to the main sponsor of the Mobility Data Contest - Google - we have three great prizes for the first three places. Of course, we cannot tell you now what you can win - it's a surprise! The only info we can share with you now is that all three prizes are splendid!

Examples

If you need some inspiration for cool mobility experiments, below are some:

Can you walk in a straight line being blindfolded and deaf? - This is a well-known phenomenon (more info here: http://vimeo.com/17083789), which still lacks a convincing explanation. Is there any correlation between being right-handed and the shape of your trajectory? Can a group of cooperating blindfolded and deaf people walk straight?

Do locations have something in common? - From the mobility perspective some locations are very similar. For instance, different trajectories, which have the same origin and destination points, cross some specific, distant locations. Can you spot such locations in a city? Can you think of some applications, which would rely explicitly on the characterisation of such locations?

Where is my familiar stranger? A familiar stranger is a person who is recognized from regular activities (commuting to work, dining in a restaurant, etc.), but with whom one does not interact. Can you find a familiar stranger among the conference participants or Washington DC inhabitants? Where do you meet them? In what context?'