We need data for @matchFWD.

A lot.

We need data to help us suggest the best  talent matches for an open job opportunity to our connectors.

We’re only 2 developers, and I spend too much time not coding so we need help.

There’s currently 2 API that buzzed trying to evaluate influence, reach, reputation and authority of Twitter users : Klout and PeerIndex.

I am not really interested by most metrics right now, but I want to know what topics a user tweet about. Lucky for me both API have such a method so I ran a small manual test on Twitter users I know and follow well enough to judge the results

Here are the results…

Ben Yoskovitz (byosko)

  • Klout : ”Entrepreneurship”,”social media”,”Human Resources”
  • PeerIndex : “private equity”,”venture capital”,”finance”,”marketing”,”montreal”


  • Klout : “Montreal”,”Yul”,”Developers”
  • PeerIndex : “free software”,”realtime web”,”google chrome”,”science”,”api”

Carl Mercier (cmercier)

  • Klout : “Montreal”,”Met”,”technology”
  • PeerIndex : “databases”,”data management”,”digital media”,”ipad”,”iphone”


  • Klout : nothing….
  • PeerIndex : “cycling”,”andy schleck”,”bessette”,”bicycle sharing system”,”cervlo”

Martin Dufort (mdufort)

  • Klout : “iPhone”,”Apple”,”Google”
  • PeerIndex : “iphone”,”cloud clients”,”embedded linux”,”apple tv”,”nexus”

Seb Provencher (sebprovencher)

  • Klout : ”"Montreal”,”Marketing”,”Quebec”
  • PeerIndex : “marketing”,”mass media”,”realtime web”,”digital media”,”bing”

Alex Dao (alexdao)

  • Klout : “Montreal”,”Retailers”,”Regions”
  • PeerIndex : “montreal”,”realtime web”,”google search”,”marketing”,”villemarie montreal”


  • Klout : “Quebec”,”Techno”,”Francophone”
  • PeerIndex : “mass media”,”collaboration”,”marketing”,”community websites”,”criticism of facebook”

Klout seems to limit the number of topics to three and is much more generic. If I had to classify my users in large bucket, that could be interesting but it wouldn’t make it easier to identify specifically everyone in my twitter feed.

PeerIndex results are much more precise and diverse. The downside, I would guess without testing, is that they are probably much more volatile. The topics you rank for this week, won’t be next week topics… So how do you deal with this… You weight all tags you attach to your users ?

What am I going to use ? Most probably both. Now let’s it Stackoverflow API and GitHub API to find great developers…

Do you use these ? How ?

Enhanced by Zemanta