Comments on: Column: Chartering a Course Through Data http://annarborchronicle.com/2010/02/22/column-chartering-a-course-through-data/?utm_source=rss&utm_medium=rss&utm_campaign=column-chartering-a-course-through-data it's like being there Tue, 16 Sep 2014 04:56:38 +0000 hourly 1 http://wordpress.org/?v=3.5.2 By: Rod Johnson http://annarborchronicle.com/2010/02/22/column-chartering-a-course-through-data/comment-page-1/#comment-40343 Rod Johnson Wed, 24 Feb 2010 14:09:12 +0000 http://annarborchronicle.com/?p=38208#comment-40343 It made me laugh that there’s an apparent drug bust in that hotbed of narcotics activity, Loch Alpine. I think Google Maps got confused by “S WAGNER OR PARK LAKE AVE”–that’s probably supposed to be Dolph Park.

]]>
By: Edward Vielmetti http://annarborchronicle.com/2010/02/22/column-chartering-a-course-through-data/comment-page-1/#comment-40328 Edward Vielmetti Wed, 24 Feb 2010 06:36:37 +0000 http://annarborchronicle.com/?p=38208#comment-40328 I went back over Dave’s sample data set and didn’t see Plymouth Road thousand blocks, so it’s not on this particular map; I was able to confirm, though, that all but one of the 1548 records in the data set available tonight at http://a2gov.org/data that includes the word BLOCK that have a 4 digit address in front of the block have either 1000, 2000, or 3000. The odd exception is a single call to an Ypsilanti address in the 7000s.

]]>
By: Bob Martel http://annarborchronicle.com/2010/02/22/column-chartering-a-course-through-data/comment-page-1/#comment-40313 Bob Martel Tue, 23 Feb 2010 22:20:58 +0000 http://annarborchronicle.com/?p=38208#comment-40313 The “hundred block” approach to providing the crime statistics probably provides some form of privacy for the victims. I’m not sure I’d want the whole world to know that my particular house was broken into. That might give others some ideas. The “thousand block” limitation as highlighted on Plymouth Road might be a bit of over kill if victim privacy is the intention.

]]>
By: Joe Hood http://annarborchronicle.com/2010/02/22/column-chartering-a-course-through-data/comment-page-1/#comment-40283 Joe Hood Tue, 23 Feb 2010 05:36:07 +0000 http://annarborchronicle.com/?p=38208#comment-40283 Curious that the police still use the hundred block for location. Who does that benefit? I guess this is the way that police have always done this and only now that citizens get the data is there a need to change. Does the Ann Arbor Observer need to cross check all of the police reports with reality (sounds like a hell of a service on their part–which should be unnecessary).

Thanks for another well reported out article (I guess the IT budget that a2politco.com rails about has benefits).

]]>
By: Dr Data http://annarborchronicle.com/2010/02/22/column-chartering-a-course-through-data/comment-page-1/#comment-40279 Dr Data Tue, 23 Feb 2010 00:50:17 +0000 http://annarborchronicle.com/?p=38208#comment-40279 This article is a very nice explanation of the difference between reports and raw data. Raw data are more flexible, but not everyone will be able to make use of it. For those folks, the monthly crime map in the Ann Arbor Observer is a useful alternative. Otherwise, an interested citizen might have to find someone to help them massage the data, but that’s preferable to waiting for a FOIA request.

This article was easily worth my monthly volunteer subscription.

]]>
By: Edward Vielmetti http://annarborchronicle.com/2010/02/22/column-chartering-a-course-through-data/comment-page-1/#comment-40275 Edward Vielmetti Mon, 22 Feb 2010 22:21:40 +0000 http://annarborchronicle.com/?p=38208#comment-40275 A note for the mapping project for the police service calls -

The data is mapped at the block level, so “400 block of Mushroom Court” might mean 401, 435, or 499.

The dataset I saw mapped street numbers greater than 1000 into blocks that were 1000 big, not 100 big, so “2000 block of Washtenaw” might be 2001 and it might be 2999, a mile or so away.

The visual impact of this is to make downtown look like it has more crime, and the outsides of town look like they have less crime; that’s an artifact of the coding used, not an accurate portrayal of reality. You can see this most clearly in northeast Ann Arbor where there are way fewer records along Plymouth Rd visible (since they are all bunched up at 2000 plymouth)

A plausible remedy to this would be to randomize the address mapped; so when you got a record that said 400 block of Mushroom Court you’d randomly pick 435 some day and 453 some other day. It’s more useful along blocks that are bigger.

]]>