bandicoot, an open-source Python toolbox to analyze mobile phone metadata

Easy-to-use environment

bandicoot provides a complete, easy-to-use environment for data-scientist to analyze mobile phone metadata. With only a few lines of code, load your datasets, visualize the data, perform analyses, and export the results.

1400+ behavioral indicators

bandicoot indicators fall into three categories: individual (e.g. number of calls, text response rate), spatial (e.g. radius of gyration, entropy of places), and social network (e.g. clustering coefficient).

Install bandicoot

bandicoot is on Github and PyPI, and it's really easy!

pip install bandicoot
python
>>> import bandicoot as bc

Data visualization

bandicoot has built-in visualization tools. Load a user's file and visualize his social graph, mobility pattern, and interactions.

Check out our IPython notebook for live examples.

Data verification

We detect and warn you of potential missing data (no location, wrong date…). bandicoot automatically exports more than 40 reporting metrics to help you detect issues.

Research papers using bandicoot

If you use bandicoot in your research please cite it as: de Montjoye, Y. A., Rocher, L., & Pentland, A. S. (2016). bandicoot: a Python Toolbox for Mobile Phone Metadata. Journal of Machine Learning Research, 17(175), 1-5.

Yves-Alexandre de Montjoye

Imperial College London

Luc Rocher

Imperial College London

Alex ‘Sandy’ Pentland

MIT Media Lab

As well as Florent Robic, Kevin Mustelier, Walter Menendez, Zachary Neeley, Keeley Erhardt, William Navarre, Brian Sweatt, and Bjarke Felbo.

Sign-up to our newsletter to receive updates (2-3 times a year) about bandicoot.

Contact

Feel free to ask questions and report issues on our GitHub page or to contact us at X@Y where X=demontjoye, Y=imperial.ac.uk.

bandicoot