Introduction: Wifi for Social Good

Humans are decently good at recognizing patterns in language and in numbers – we’re good at identifying patterns in data. But what happens with a new source of information?

Take an example piece of data from our project. What could the following message mean?

00-14-22-03-26-12 graduate-student computer-science

This is a person’s use of a GTwifi router, who happens to be a graduate student in CS. The leading numbers in this message represent a wireless internet access point tied to a specific location on campus. GT has hundreds, if not thousands, of these devices so that users can seamlessly maintain WiFi access as they move across campus. No one really notices them or thinks of them as data points.

As data scientists, it’s our task to recognize potential sources of data in these often unconsidered areas of our lives. We also digest and present information to others in meaningful ways. But before we can start to analyze data, we have to figure out the question itself. This is what our group is thinking about this week – what do we do with data like these WiFi access records?

A a project group last year focused on determining when users were present in a specific context (i.e. in the Starbucks line), our approach can be general and holistic. With information about how many students of a specific specialization are present in each of our campus buildings, we can assess migration patterns between these nodes, as well as who may use buses to move across campus.

Consider the following scenario: Building A is currently filled with students of various kinds, and is located very close to Bus-stop A. One might expect that, when the steam whistle signifies the next round of classes, Bus-stop A will be overloaded with students from Building A. However, what would happen Building A contains 90% A-Students and only 10% B-Students, who are the ones that need to migrate to the B Buildings across campus? The nearby bus stop would not see much traffic at all:

a and b buildings
The new missionaries and cannibals problem. (Which Georgia Tech department represents the cannibals is omitted for now)

Ideally, analyzing and visualizing patterns such as these can help us understand how students and faculty not only migrate around campus, but also interact with connected artifacts such as GPS-located busses. Outside of this Georgia Tech scope, the implications are even more numerous, and our later posts will have much more to say in that regard.