Using Skyline for detecting anomalies in Congressmen's voting patterns

[Vitor Baptista]

Hi,

I'm researching about how to better monitor the Brazilian congress. My hypothesis is that it's possible to notice changes in the congress through changes in the behavior of congressmen. For example, if the congressmen of a party (there're many parties in Brazil) used to vote against the government in nominal votes, but then start voting with them, could mean that an alliance was forged. Maybe it could be correlated with the nomination of the leader of that party to some ministery, for example.

Right now I'm trying to use tools like Graphite and Skyline to track these timeseries (voting patterns, changes in government, nominations, etc.). The first issue I had was that these tools were designed for monitoring much bigger data, often with metrics gathered per second. All timeseries that I'm using have a much slower frequency. Voting, for example, are in the hundreds per year. By tweaking some configuration values, I was able to make Skyline run and detect some anomalies, but there's still much that will need to change, specially on the algorithms side.

I'm wondering if someone knows any anomaly detection tool or research that uses small data on a bigger time scale like this.

Cheers,

Vítor.

[Abe Stanway]

I wouldn't recommend Skyline for that purpose - you're probably much better off using R. If you specifically want Skyline's algorithms, you can use Crucible (github.com/astanway/crucible), or just take them out of algorithms.py and build your own analysis workflow on top of them.

--
You received this message because you are subscribed to the Google Groups "skyline-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to skyline-dev...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Abe Stanway

abe.is

[Vitor Baptista]

Hi Abe,

I was trying to use Skyline for the visualization, but maybe I'll be better off building something from scratch. I'll take a look into Crucible as well.

Thanks for the quick response.

Cheers!

--

Vítor Baptista

Developer  |  http://vitorbaptista.com | LinkedIn | @vitorbaptista

The Open Knowledge Foundation

Empowering through Open Knowledge

http://okfn.org/  |  @okfn  |  OKF on Facebook  |  Blog  |  Newsletter