I wanted to prove that favoriting semi-random tweets increases followers in Twitter. I wrote a “bot” that favorites tweets that contain hashtags that I’m interested in. For my own peace of mind, I didn’t choose mainstream or popular hashtags (ie. #KathNiel) because I wanted to have a set of quality followers. I’m sure there have been experiments like this before but I pursued this because:
- I wanted to practice basic Ruby syntax
- I wanted to try Sinatra.
- I was bored. (Mainly because of this)
I specifically chose the following hashtags which are relevant to me:
I don’t want any useless people following me anyway.
The Script (Nope, not the band)
I made this very simple script: https://github.com/katpadi/xdaysb4nba/blob/master/fave.rb. This script does the “favoriting”. I created an app first in Twitter dev site for this script to work and get tokens.
So what the fave.rb does are:
- Search tweets containing the hashtags above.
- Pick a random tweet using the Ruby method sample.
- Favorite it and log.
The second part is the scheduler. The script just uses a basic cron script here: https://gist.github.com/katpadi/425ea20db85d141e432. It makes the fave.rb script run 24 times a day at random times.
Results (so far)
I started running the script Sept 18 and so far after ~9 days. My followers count increased somehow.
Before I started the twitter favoriter experiment, my followers count was at 472. By now it’s 486. The peak (so far) is 486. Note that some users unfollow randomly. They might be bots or I don’t know.
Another observation: I’m kind of sure that the followers were gained from the “favorite” act because as I checked some of them, they are either basketball fans or devs.
Here are all the favorited tweets: http://katpadi-fave-experiment.herokuapp.com/P. S. (Not related & relevant to the experiment at all) — I was just making an auto-tweet countdown for Twitter before but this “favorites” thing got me hooked in so I decided to do it. That’s why it’s in the same repo. (Messy, I know :p)