MIT | ~ Mark's Musings ~ blog.markwcarbone.ca

As we spend more and more time online, one leaves a digital footprint – digital connections to our name which are part of your online presence.

Personas is a component of the metropathologies, currently on display at the MIT museum by the Social Media Group. It uses sophisticated natural language processing and the Internet to create a data portrait of one’s aggregated online personaliy. In effect, Personas shows you how the Internet sees you.

Here’s how it works: You enter your name, first and last only, no middle initials, titles etc. and Personas will dig through the Internet to search for information and attempts to characterize the person by fitting the information against a predetermined set of categories that an algorithmic process created from the total data set found. Each step of the multi-stage process is visualized. The final result is the representation of the seemingly authoritative personal ‘online’ profile (based on the data set).

Personas is a great example of data mining to build profiles from extensive quantities of information. As amazing as this process is, it is also an example showing inconsistencies and inadvertent errors that occur, mainly due to the lack of ability to separate data from multiple owners of the same name. In other words, if I run the process using ‘Mark Carbone’ the final profile presented will be based on the collective data for all of the people with the name ‘Mark Carbone’, not just me.

The algorithm fits the collected data against 23 categories, so it’s fairly comprehensive. The categories include one called ‘illegal’ and I would bet that this would make many of us squirm a little … I road tested a few names, including my own and this category exists for all the names I tested. None the less, this is a great example a powerful data mining tool.

As an example the representation of ‘Mark Carbone’ looks like

personas_example