Most recently updated graph.
Russian version.

Step 1 - Scraping the users I follow

Use Selenium to get the list of people I follow

Step 2 - Scraping the users they follow

For each user repeat the cycle of Step 1. In addition for future Machine Learning research - record the originating user

Step 3 - Scraping account info

For each user get account information and front page information for Machine Learning models and core bot functionality

Step 4 - Prepare information for bot activity

Generate the tables that bot uses for following users and liking their posts

Step 5 - Start the bot

Simply clicking "follow" and liking photos with random delays and longer delays for "follow", some accounts take longer to load.

Step 6 - Next day check and cycle repeat

Since the set was split in three daily chunks, this step repeats step 5 and checks my account info for the people who started following me.

Step 7 - Unfollow

Clean up the account so that I can get back to following people I did before, in normal mode.

Step 8 - Compare the lists of followers

Generate the table of the followers for the graph. Compare the original list to the cumulative sum of three days of bot activity and get the new list of followers to see who stopped following.

Step 9 - Graph

Generate a custom graph using p5.js library

Next Steps