A couple of weeks ago I wrote about a project I was working on called City-Color. I'm afraid to say there is nothing to update on that, except that I got lost in side projects until my desire to see the main project come to life was drowned.
Along the way I started using twitter again but since the last time I really used twitter was in 2009 I needed to redefine who I follow. I had a head start thanks to a sweet blogpost I had stumbled onto with a graph of influential data science twitter accounts a few weeks back, but this only helped me flush out my data science tweeps, I still needed to find people to follow twitting about node.js and poetry. And so I started the social discovery process.
- Search for a topic and follow people with about-me's that seem to fit the subject I'm looking into.
- Follow accounts mentioned by accounts recently followed that I'm enjoying.
- Follow accounts retweeted by accounts recently followed that I'm enjoying.
- Redact my list, unfollowing accounts I deem relativity inactive or inane or redundant (aka: RTing other account I follow).
Though I think this method is rather sound (in just a couple of weeks I'm managed to turn my twitter stream into a nearly constant stream of content I want to read) it defiantly lends itself to a progressive deviation from the topic I searched for in step 1.
Twitter friends don't always work in the same field as twitter friends, and while step four helps weed these accounts out, I can't help but wonder if there is another community of users I'm not finding. Of course, there are other methods, (I've found the "Suggestions similar to _____" Twitter email to be helpful), but these methods tend to be passive when I am actively trying to find new accounts to follow. So I asked myself:
Q: What is it I want from someone I follow.
A: Links to content I want to read.
This is what I think twitter is great at. Yes, its fun to yap back and forth with people that share your interests, but I think this is second (by far) to finding long form content worth my time.
Thanks to a gist I put together to simplify tweeting quotes from blogposts, all the links I post to Twitter are saved in a Bitly account. So I started looking through the Bitly API, asking questions about how I could take this data and turn it into a tool to help me find twitter users I'd want to follow. There were a lot of side projects and a few times I thought this project too would be drowned in my recursive learning style but thanks to Mark Starkmans's blogpost I was able to take a step back and ask myself:
Q: What is this project for? Shipping or Learning?
A: Shipping, I want to know who tweets the links I bitmark.
Q: Does this Google Chart of average click rate solve that problem?
A: No....
So that I can now present to you Bitly Fellows. A single page Bitly App that takes a username and returns a list of twitter accounts ranked by the number of your bitmarks they tweeted.