Mood: curious
Posted on 2007-03-29 15:41:00
Tags: haskell congressvotes
Words: 166

Allergies seem to be on the retreat, thanks to a little extra sleep today and tons of tea and orange juice and hot chocolate.

I should have worked on congressvotes yesterday, but I didn't have the activation energy. The problem is that parsing the XML and extracting any sort of data gets really really really slow - like 37 minutes to do a little bit of stuff to all 600ish votes from 2006. I've looked at the code and I can't find anything in particular that's inefficient about it, but I'm pretty new at Haskell. The next strategy is to do an offline transform to some sort of friendlier text format with just the information I need, and hopefully the web service, etc. can just use those text files. Hopefully I'll give this a shot sometime soon. (for some reason this post about premature optimization spoke to me; it's a good general philosophy!)

Wow, are there really going to be Kwik-E-Marts? I could go for some Krusty-O's...


Comment from tehfanboi:

So as someone who has a degree in political science, I never remember doing this much science in any of my classes. But I am curious where this is headed. Like possible applications? Mostly curious.

Comment from gregstoll:

The original idea was to choose how you would vote on certain bills and then see how similar your representative is to you. Kinda like one of those OKCupid thingies. (maybe selecting "important" votes from what various interest groups think is important enough to rate on their annual scorecard)

But, once I have the data we could do all sorts of fun stuff, like finding the most contrarian representative, etc. I'm open to ideas!

Comment from wonderjess:

my friend emily reads my friends page sometime and read about your project...she's a second year poli sci phd student up at harvard and she says she has lots of data for you if you want it. :)

Comment from gregstoll:

I heart data! Have her send me an email if she doesn't mind :-)

Comment from anonymous:

I have never had cause to do performance optimization any code I've ever written, ever. Which isn't bad, because it frees me to think about how to architect my code for maximum maintainability, but it makes my game programming classes an interesting experience, because performance is almost all my teachers think about.

Comment from brittongregory:

Whoops -- that last one was me. And let me amend that statement to "almost never", and it's certainly rarely come up before actually testing the code.

Comment from anonymous:



This backup was done by LJBackup.