05 Jan 09 Not so surprising, aggregators lead in URL scoring

I thought I’d see which Twitter users are scoring the highest in terms of posting URLs that become popular. My code gives them points based on how early they posted and how popular the URL becomes. I suppose it should not have surprised me to find that most of the high scoring users are not real people, but aggregators that feed tons of URLs.

Who is it that says that web analytics data is always messy? Whoever it is, right you are! Since a fundamental goal of the work I’m doing is to uncover interesting points of view, I need to downgrade sources that aren’t behaving as though they really have a point of view (or at least an intelligent one). I can tell instantly that I’m almost certainly looking at an automated system when I see that the “user” in question follows zero or very few people. That’s grounds for immediately downgrading. I’m not sure if I want to downgrade based on the volume of postings. Certainly beyond a believable number… and perhaps if every single post contains a URL.

Here are the top 20 sources from the last week or so, based on the criteria I described above.

1. Net2 (878)
2. techupdates (706)
3. OriginalSignal (587)
4. radi8 (565)
5. Dakshinamurti (542)
6. GaryTheGeek (453)
7. techupdate (449)
8. haripakorss (436)
9. readmashcrunch (392)
10. twittfeed (379)
11. TwitLinksRSS (359)
12. top_post (342)
13. tclauss (329)
14. TechFeed (303)
15. tc2tw (300)
16. vcsangels (295)
17. dlbrown06 (287)
18. davidsim (279)
19. mashable (272)
20. ReTweetTrends (268)
21. balduaashish (268)
22. wiredgnome (264)
23. julieti (259)
24. TechRSS (248)
25. davekresta_rss (246)

Tags: social network analysis, twitter