Funf is an open source framework for collecting and analyzing mobile data. It has been used by MIT to see how political opinions change during an election campaign, how users interact with each other, or how illnesses spread through population.
The MIT Media Lab has created the Funf Open Sensing Framework, an open source sensing and data processing framework for mobile devices. The basic idea is to install a collector application on mobile phones, tracking all sorts of activities performed by a group of users, activities that are later analyzed in order to determine certain patterns in users’ behavior. The MIT team used the software “drawing surprising conclusions about the way political opinions, dietary habits and illnesses — among other things — spread through populations”, and the framework can be extended to analyze all sorts of activities.
The basic concept used by Funf is the probe which is a software module collecting data offered by the low level phones sensors (accelerometer, gyroscope, proximity sensors, temperature sensors, etc.), but there are also probes doing higher level data collecting such as the “activity monitor”:
The “activity monitor” probe, for instance, can distinguish the accelerometer data typical of, say, a phone on the hip of someone being jostled on a subway train from the data produced when the same person is walking briskly or climbing stairs. It can thus provide a single numerical score for the user’s physical activity over any specified time span.
Funf comes with about 30 built-in probes related to device, position, motion, environment, device interaction –running apps, screen activity, browser bookmarks, etc.- and social –contact information, call log, SMS. The framework can be extended to monitor other types of activities.
For protections, data collected is encrypted, and sensitive data, such as contacts or SMS text, is hashed. The user also has the option to remain anonymous by having the data collected not related to himself.
The source code of the framework is available on Google Code along with samples of using it. The code has two main components: the collector, an application that is installed on an Android phone, and a number of scripts for unpacking the data collected into an SQLite database and for visualizing the data. An API makes it possible to integrate Funf functionality in other Android applications.
Funf has been open sourced under the LGPL license and it is sponsored by Google, Samsung and Motorola.
The Wall Street Journal published an interesting article last year mentioning some of the current projects (including Funf) attempting to analyze human behavior and interactions using mobile phones, the results achieved so far, and the privacy implications of such projects.