Comments | pmb: Estimating homelessness via too-clever thinking and long train rides

pmb

Estimating homelessness via too-clever thinking and long train rides

Mar 14, 2006 02:12

Grar. Too much to do and think about and too much too much too much. So here's an idea I had on the train this weekend...

While riding the train, sometimes I play the game 'where would I sleep if I were homeless?' I've recently started playing a variant called 'How many of those spots are already taken?'

The resulting number is far too high. ( Read more... )

Comments 12

(The comment has been removed)

pmb March 14 2006, 19:43:19 UTC

It does suggest that, but later comments indicate that we might want to scale it down a bit. I'd back 'at least 2x'.

freyley March 14 2006, 17:53:24 UTC

I don't follow the assumption that homelessness grows with quadratically as you expand area. If the area following the train line is 4x as dense in one place, you would expect the city to be 4x as dense. This should follow the same logic as if you catch-tag-and-release birds and you get four times as many non-tagged as tagged, there are 5x birds.

Except that I would suggest that trains and train lines are likely to be _more_ densely populated by homeless people, because a) they are a source of transportation for some percentage of the homeless (granted, not a huge percentage any more), and more importantly, b) they provide better shelter. The latter is more important and arises out of two factors: 1) trains more often go under roads than over them, as they were built first, offering more bridges to sleep under, and 2) trains don't stop. Cars can. It's harder for train police to find you than regular police. Train bridges are therefore way safer.

freyley March 14 2006, 17:56:39 UTC

Oh, wait, I misread. 4x farther out of town would make sense with 16x homeless population, except that towns aren't circular and trains tend to create population centers (and population centers tend to follow trains), at least old centers did. Older centers are likely to be more livable for homeless people, as they're less likely to be heavily suburban. So now I have no idea how to read 4x farther out of town in terms of a population boost.

freyley March 14 2006, 17:57:32 UTC

But it seems like a fun challenge, to figure out how it should apply. Thanks!

And now for something completely different... canarasekal March 14 2006, 23:48:09 UTC

Hey Peter!
What's with the bags under the eyes?
- A concerned parent

Re: And now for something completely different... pmb March 15 2006, 03:40:50 UTC

It was a long tiring week and weekend and the picture was taken at the end of a long day in the long weekend. They are not a permanent feature, I was merely extremely tired when the picture was taken. Your kid is still okay.

Re: And now for something completely different... kuddliphish March 15 2006, 13:04:33 UTC

Dude, I assumed that it wasn't a picture of you because it looks like you've got, you know, facial hair. Have you really grown a beard? Is that even possible?

Re: And now for something completely different... pmb March 15 2006, 19:41:44 UTC

( ... )

Thread 5

FAI? Hmmm... akjdg March 15 2006, 06:37:48 UTC

A fascinatingly simple and intriguing attempt at a hard-to-crack statistic. Controlling for the geographic oddities of any specific metro area and all of the other factors listed above may prove to be so assumption-riddled as to make the result less than satisifying, but I like it anyways.

Extending it to Fairbanks (do you really mean the one in AK?) could pose some unique challenges.

1) Seasonality. How many homeless live outside when its -40F? I imagine the answer is not zero, but is also not equal to the summer population.

2) Just ain't urban enough. I've never approached FBX via train, but I'm thinking the 'urban' - rural transition is rather more distinct than in areas with more people to throw around.

3) Harder to define 'homeless', because it's harder to define 'home'. Where exactly is that shifting line between 'rustic Alaskan retreat', 'cabin in progress', look at me plywood and blue tarp castle, Arrrr', and 'homeless camp'? We were driving around Sutton a few weeks ago, and the line sure as hell blurs out there.

Re: FAI? Hmmm... pmb March 15 2006, 07:27:39 UTC

I did mean Fairbanks, AK. I always assumed that the homeless population at -40 degrees was pretty much zero.

Re: FAI? Hmmm... coldtortuga March 15 2006, 17:41:54 UTC

Huh -- well, crumb! Yesterday I commented on homelessness in Fairbanks (I believe it would have been one of the first comments on your post) but apparently its bits have been devoured.

There are homeless people in Fairbanks. As one would expect, the homeless population varies seasonally, dropping to nearly zero in winter.

As akjdg points out, Fairbanks is only urban in comparison with the rest of Alaska. For a sense of scale: the Fairbanks North Star Borough has a population under 100K but is significantly larger than Massachusetts; I just turned on Google Earth and you can see pretty much all of "downtown" Fairbanks from 7000 feet (it covers on the order of 1 square mile). I can't recall seeing panhandlers in Fairbanks, though I grew up outside of "downtown".

There is a much much bigger population of home-ful alcoholics in Alaska -- waaaaaay bigger.