@rrhoover I kind of wish it operated on ALL of my followers, not just 10K. Makes the value of the experience a little more limited, but I get they have to try to convert users to customers some how.
Hi! I'm one of the developers with Affinio. We're really excited to be launching Discovery; it's a quick/lightweight version of our full segmentation tool. The hope is to be able to analyze a user's following and use machine learning to figure out how their followers are distinct (if your followers all had a house party, who would be hanging out with whom?).
Hope you enjoy it! Happy to answer any / all questions about the product.
Oh, and thanks so much to @byosko for featuring us!
Report
@phil_renaud really cool stuff. i'm impressed. Can you tell me a lil more about that crazy colored dot D3 visualization? Are all those nodes 1 to 1 real followers? and what do the connected lines mean, that they follow that other person? Really cool
@drawby Hi @drawby, glad you liked it!
The network graph visualization is representational, but not 1:1 (our graphs have a max node count anywhere from 300 to 1500, depending on your device type). The connected lines represent, between groups, shared influencers; and within a group, likelihood of a given user knowing other users. It's a simpler model than we use when generating graphs for affinio.com (which can have up to 20 groups), but it tells the same sort of story.
This is really cool! I have just started becoming more active on twitter and figuring out my brand so to speak so this is really helpful. Definitely recommending!
@joshdance Hi Joshua - thanks for trying it out! We opted not to show the users themselves and instead show their common influencers to keep Discovery lean. However, it's worth mentioning that in the full product at affinio.com , much more follower info is available (in aggregate, including content trends, etc.)
Hi! I'm one of the developers with Affinio. We're really excited to be launching Discovery; it's a quick/lightweight version of our full segmentation tool. The hope is to be able to analyze a user's following and use machine learning to figure out how their followers are distinct (if your followers all had a house party, who would be hanging out with whom?).
Hope you enjoy it! Happy to answer any / all questions about the product.
Product Hunt
Raycast
Product Hunt
Jupe
Discovery by Affinio
Discovery by Affinio
Discovery by Affinio
Blab
Summer Bod 2020
Discovery by Affinio