The recent investigations into the practices of Cambridge Analytica has exposed how Facebook data can be used to influence elections and referenda.
CA harvested user’s likes, derived personality types though “psychographic analysis” and then targeted these small segments of users with highly tailored advertising.
Your entire history can be downloaded from Facebook, provided in an HTML format, and it is no doubt a confronting experience to see the quantity and intimacy of the information.
Most importantly though, while FB allows you to download your data, they don’t provide the insights they they can extract from it, which is what makes it powerful.
After downloading, we will start by converting your data into a more useable format, then perform some basic analysis. First think about what questions you might like to ask to the data set.
- Who are the friends I have deleted?
- How often do I swear?
- How many photos contain a bicycle? a selfie? a drink?
- Show me all the noses.
- How often did I talk to people I don’t like?
While statistics and data-visualisation will be interesting (and funny), we would like to go beyond this analysis to create an interface to deeper aspects of your own personal archive.
You will examine your data in detail, refine it, augment it with other sources of information in order distill new insights into yourselves or the broader social media ecosystems.
Whether through classification, manipulation, or generation; your interface should document of expose these ecosystems.
Your interface might also respond to input. Consider what inputs you could use to extract these insights: speech, gestures, text, hardware values, date ranges.
As a group, consider how these interfaces might communicate with each other. Could their inputs and outputs be combined to show something unexpected?
At the end of the week we will present an installation in the studio.
While we’re using Facebook data for this workshop, the skills we will cover can easily be applied to other data sets.
- Primary languages: Node, JavaScript
- Secondary languages: Python, PHP
- Data parsing
- APIs
- Basic NLP, POS tagging
- Layout: HTML, CSS
- Communication: WebSockets, WebRTC
- Speech-to-text, image tagging, facial recognition