Facebook datamining

How a random weekend project turned into my graduation work… In the beginning, it was just that. I wanted to learn how to work with databases. I figured that instead of using some random JSON example I could download my data from Facebook and use that instead.

I’m already satisfied with the basic functionality (the original objective) - the Python script counts the number of messages in a specified time frame, say a week, resulting in messages per week value which it then puts into a .csv chart. I can then take this file, open it in Excel and make a nice graph out of it.

I recently also implemented the most used word counter. I’m planning to add a couple more features and make the script more interactive and intuitive to use. I haven’t yet figured out whether this data could be used for at least rudimentary psychological research.

Update: I refined the project and made it public. It’s certainly not finished yet, but it works, for my use case at least. If you clone it, it should run just fine, it uses just one non-standard package (tqdm to show progress bars) but you can edit that out of the code, it’s not mandatory (just don’t freak out when it hangs for a couple of minutes, sometimes it does take quite a while to chew through all those messages).