The science of physical geography – of surroundings we can sense, visualize, and analyze – is one of static, measurable characteristics. We gauge elevation with altimeters & GPS, roads & rivers can be seen in satellite imagery, we can even identify buildings and count vehicles in a parking lot. Much about physical geography changes slowly, and in cases when it doesn’t, we can readily observe changes to correct error. Socio-cultural elements, on the other hand are fluid, dynamic, elusive – we can’t turn to a handy piece of scientific equipment to tell us an answer. More novel approaches are required to measuring the socio-cultural dynamics of places.

It’s difficult to fit restrictive definitions of socio-cultural fabric into a set schema. Defining the pattern-of-life for a particular social group or the people of a certain region can be nearly impossible, but we can get close by producing models or formulas that can extract greater meaning from esoteric data.

Baghdad Alpha
(Photo: Aaron Straup Cope)

Modeling the dynamics of human behavior is challenging, so we need mathematical and statistical frameworks to digest and understand waves in the cultural fabric. Some time back, the folks at Flickr released the tools to produce alpha-shapes from geotagged photos on Flickr. The result is a natural, fluid depiction of “neighborhoods” as defined by the GeoPlanet spatial hierarchy. Neighborhoods are not bounded by the rigid bounds of streets, rail lines, and buildings. A person’s perception of their “neighborhood” consists of intangibles like ethnicity, demographics, lifestyle, and language. Alpha shapes are mathematical models that can help us to visualize boundaries between points of non-physical data, like a person’s “influence” or the sentiment in an area following an election.

Eric Fischer, a map designer and developer, has done a series of data visualizations wherein he takes readily available datasets or streams – such as geotagged photos, tweets, US Census data – then extracts and fuses data in clever ways to pull out meaning not visible on the surface. By combining what people are saying (on Twitter) with what people are seeing (on Flickr), features start to climb out of the data. Which areas are people in when sharing about their social lives? Which are they more prone to want to remember with a photograph? What areas are active in both? When we look at visualizations like this in timelapse, how would the models change shape?

Tweets and photos from London (Photo: Eric Fischer)

We have to take statistics, demographics, and social graphs and merge these data with other “known” quantities we can gather, like economics, ethnography, linguistics, and history. Only with a composite of many different variables can we effectively model the constantly-changing dynamics of societies.