Stalky

This repository no longer works ever since Facebook decided to switch from HTTP requests to WebSockets

What is this?

A complete remake of a project by Alexander Hogue. Read his blog post to get an idea of what this is or watch his talk at PyCon Australia 2016.

Why is this?

As a student with the circadian rhythm probably completely out of tune, I thought it wuld be a fun eperiment to check whether other students struggle to go to sleep at reasonable times as well.

Researching the topic of sleep pattern analysis I stumbled upon this repository, which collects data from Facebook's internal API, but Facebook changed the API long ago and the program didn't work, so I decided to use the repo as a starting point in making a new tool.

Installation

Run

make install

You'll also need to supply some way of authenticating yourself to Facebook. Do this by creating a SECRETS.json file with the following fields:

{
    "uid": "<Your Facebook user id>",
    "cookie": "<Your Facebook cookie>",
    "client_id": "<Your Facebook client id>"
}

You can find your FB client ID by inspecting the GET parameters sent when your browser requests facebook.com/pull using your browser's dev tools.

Gathering data

make fetcher

This will run the fetcher script indefinitely (restarting on crashes), creating data in a SQLite database (data.db by default). You can for example host this on a microcomputer (e.g., Raspberry Pi) running 24/7. Depending on the number of Facebook friends you have, and how active they are, you can expect around 3 - 6 MB per day to be written to disk.

Plotting some graphs

1. Run make server to start the visualization webapp
2. Go to http://localhost:5001 to view it
3. Submit the FB User Name of a user whose activity you want to graph and the time span.

The graph library used is blazing fast, allowing you to graph months of data, zoom in (select a segment of the graph) and out (double click on the graph) and pan around (Shift + Click). You may want to create an index on the Logs table in the database, to further speed things up.

The "webapp" uses basic authentication that can be enabled by creating an auth_hash.txt file which contains a MD5 hash of a concatenation of valid username and password pair.

Collected Data Insights

Using data collected over almost 3 months (01.2019 – 04.2019 with a fetcher downtime in the middle of february) I was able to perform some basic data analysis. Here are the results: