Python code to explore the increase in the number of Facebook messages exchanged during full moons. I talk a bit more in detail about the phenomenon on my website.
This code makes use of the Full Moon Calendar 1900-2050 found on Kaggle.
Steps:
- Request all-time message history from facebook (low image quality is enough, only the .json files are needed)
- Depending on your past facebook activity, multiple folders will be available for download. Find the one where the JSON files are located and execute the rest of the steps in that particular one
- Run the json_keeper.py then the json_mover.py to collect all JSON files into the top-level folder. Delete all subfolders and non-JSON files.
- Run read-absolute-distance.py to get a frequency dict with the number of messages sent before or after the closest full moon by X days.
- Run read-relative-distance.py to get a frequency dict with the number of messages sent around the closest full moons for each message (+/- 3 days)
Please note: when looking at the absolute distance values, we notice a large decrease in the number of messages. This is because '0' distance means only one day is considered for the data. Other distances contain two days, example: 1 distance from full moon is the sum of messages sent 1 day before and one day after the full moon.
