MartyConnelly
martyconnelly at shaw.ca
Thu May 24 14:03:45 CDT 2007
Yes, after a lot of arguments about people no longer paying for phones, dying off, getting married and changing name by deed poll not to mention multiple people sharing a phone eg. roommates etc. I tracked internal moves but can't remember the exact number, as address matching was in it's infancy I don't trust it and I was one of the infants doing it. Every town has its quirks that can affect these types of statistics, for example in the city of Montreal, all apartment leases end May 1'st. This can skew statistics depending on the time frame, never mind trying to find a moving van around that date. Also people tend to move less often in small towns. Steve Erbach wrote: >Marty, > >Actually, based on the data I have, I was surprised that there weren't >MORE people that had moved out of Neenah. Out of 18,000 or so unique >individuals (some with multiple voting history records), only about >1100 had moved out of the city in the 3-4 year span of the data. The >rule of thumb I heard ages ago was that about 10% of the populace >moved every year. Maybe a bunch of them moved to different addresses >within the city, but there's no way to tell that since there's only a >single address for each voter...no address "history". > >Your story is very interesting. Did the local urban planners believe >you after you analyzed the Bell data? > >Steve Erbach >Neenah, WI >http://TheTownCrank.blogspot.com > > >On 5/20/07, MartyConnelly <martyconnelly at shaw.ca> wrote: > > >>Something to keep in mind when using data and addresses. >>Just as an aside to this I remember researching the number of people who had >>moved within a year in a major Canadian city by using GIS and tax roll >>assessment data. >>I came up with around 30%. This was disbelieved by local urban planners. >>I reconfirmed the number by checking the number of Bell residential >>phone line >>connects and disconnects. This was 30 years ago before the use of cell >>phones. >>This number may vary by town size and country but is still substantial. >> >>Steve Erbach wrote: >> >> >> >>>Thanks. Unfortunately it would be a story interesting only to people like us! >>> >>>The idea of a state-wide voter list isn't bad. In the lists I got on >>>Wednesday (since consolidated in Access into one list), there are >>>about 1100 names of people whose address has changed since they voted. >>>These 1100 people have moved out of the city of Neenah. The state >>>list shows their new addresses which is handy, when you think about >>>it. If I had gotten those four years of voter history from a >>>Neenah-only database. Those 1100 peoples' addresses would simply have >>>been wrong and any mailings I did to them would bounce back or get >>>forwarded, maybe. In any event those names would be useless to me if >>>I were trying to get them to vote for me. They CAN'T vote in Neenah >>>elections any longer. At least with the state list I can see that >>>they've moved and I can eliminate them from any mailings. >>> >>>Actually I'm now a bit confused. Back in December when I sat with the >>>City Clerk and looked at the on-line database system, I found that the >>>names and addresses were stored in good form. That is, the addresses >>>were separated into Pre-Street-Direction, House Number, Street name, >>>street type, Post-Street-Direction, unit number, unit type, etc. This >>>is what I'll have to do "manually" (with Access, thank god!) with the >>>names I was given since the addresses are all in one field. During my >>>re-election campaign I made lists of voters by street and by house >>>number so that I could move sequentially along a street and know which >>>homes had voters in them. The number of people who actually vote in a >>>February primary election is minscule. This is good since that was a >>>damned cold month for door-knocking. One wants to target one's >>>door-knocking as much as possible. That December list I extracted >>>saved me a lot of work. >>> >>>Therefore, I have to assume that the City Clerk chose a pre-formatted >>>report for the voter history I asked for...and that pre-formatted >>>report had several export options, one being Excel. >>> >>>Looks like I'll have to do some more digging. That is, find out more >>> >>> >>>from the state elections board or whoever has control over the >> >> >>>state-wide lists now. It would be useful to know all the options one >>>has for extracting data. I'm not certain, though, if a regular >>>citizen has access to the data. >>> >>>I'm just happy that I got these lists before the prices went up. I >>>miscounted: there were actually close to 53,000 records. $25 setup + >>>$5 per thousand equals $290 vs. the $15 I paid. I guess I can put up >>>with a bit of table manipulation and normalizing! >>> >>> >>>Steve Erbach >>>http://TheTownCrank.blogspot.com >>> >>> -- Marty Connelly Victoria, B.C. Canada