One of the criticisms of Walkscore has been that it used “as-the-crow-flies” distance calculations to amenities, so addresses got credit for access to amenities that might not actually be accessible by foot (on the other side of a freeway for example).
Well, that’s changing. There’s now a beta of a version that calculates actual walking routes to destinations. Very cool to see the results plotted out.
Happy to see my house still scores a 96 :-)
10 responses to “Walkscore Getting Smarter”
Very neat upgrades. Although I noticed that it still doesn’t recognize bridges and viaducts. It sees crossing Sullivans Gulch on 12th Ave as a loss of, then regaining of elevation, when it’s just straight across on the bridge.
My house scores a 67. But it’s actually a lot worse than that. I get 18.9 out of a possible 20 for “groceries” based on the presence of a tofu restaurant up the street, .3 miles away. Any actual grocery shopping is nearly a mile away at the local Fred Meyer.
I also got 12.9 out of 15 for coffee, based on a coffee place that I actually did frequent until it closed a couple months ago. My nearest local coffeehouse is now twice as far.
Why are 6 points possible for “books”? Why are “books” not subsumed with “shopping,” particularly in an era when books are rapidly being displaced by ebooks and a lot of dead-tree book sales are now done on-line and delivered.
I also have a beef with “entertainment” in that it appears to only consider the nearest entertainment destination, without regard to whether you actually use it. There’s a world of difference between having one movie theater nearby, and having a choice of movie theaters, live theater, night clubs, a concert hall, and various other destinations.
Oh, and for banking, I got 2.6 based on the location of an ATM that’s not even in my network on a college campus nearly 3/4 of a mile away.
This system needs to get a LOT smarter.
It seems to consider a nearby convenience store just as good as a (two blocks further away) Safeway.
In my case, it’s biased toward traveling northwest. It missed Lents Park, 1.5 blocks from my house, but included Essex Park – 1/2 a mile away across 82nd.
It’s similar with the original walkscore. For some reason, it doesn’t think people travel south. Destinations in Lents Town Center are omitted, while destinations just as far away on Powell and 82nd are favored.
That’s interesting. My place here in SE (Creston-Kenilworth) actually scores higher on the Street Smart (77) than it did on the ‘regular’ Walk Score (74). Explanation for that?
I see it works much better with an old New Jersey address of mine, which is your typical “scores unrealistically high because there’s a dozen Big Boxes and strip malls across an 8-lane highway” kind of ‘neighborhood’. For that address, 36 (Car Dependent) on the StreetSmart is much more realistic than its 51 (Somewhat Walkable) WalkScore.
Cora – I didn’t notice that before, but went back and checked. Same here. I have things up at Clinton & 25th in my score, but nothing down along Woodstock between 41st & 46th? Same distance, and some things along Woodstock are even a block or two closer (and a door-to-door bus ride on the 75). Hmm.
My almost outer east-side house in Foster-Powell had a respectable 86 walk score with the traditional system, and it explodes to a 94 “Walkers Paradise” with the new approach.
It really does need some way of differentiating the quality of destinations however, as my score is extra high because of all the nearby dive bars :)
I’d like to know why my home in Tigard is dinged for “number of intersections” – when I have clear straight-shot paths to anywhere I need to walk to. I have a straight access path to downtown Tigard without even having to cross Highway 99W (I can use the Tigard Avenue underpass); or use any one of a number of signallized crossings if need by (including one pedestrian-only signal).
It also eliminates one major grocery store, ignores the big huge elementary school that my son walks to that is just one block away as well as my daughter’s day care center two buildings away, and a 7-Eleven that while isn’t a grocery store, can be a convenient quick stop.
This new “walk score” seems to be biased for a Portland style grid system, while European cities have long attributed pedestrian friendliness to a lack of a grid system, and Tigard is getting penalized for having alternative side streets that have good sidewalks, and fewer interactions with motor vehicles. It just doesn’t make sense, as a pedestrian who has a vastly improved (and safe) bus stop now versus my old dreaded bus stop within Portland city limits.
Smarter, but not too smart.
I live in Lents, but…
I don’t go to Ararat Bakery for my groceries
I never take my sister to school at Village Free School
So some of the stuff on here is too lax about definitions, considering the nearest real grocery store is Fred Meyer or even 88 Market down the street. Not to a bakery. Plus my sister goes to Kelly Elementary down the road, the nearest logical school.
So yeah, needs to work out some kinks. :)
Actually, Ararat has a decent enough grocery selection. You can get baked goods, basic staples, canned goods, produce in the summer and they have a full dairy section as well as cured meats and a freezer section full of prepared foods (mostly piroshki and pelmini, but they’re good). I go there for lentils, bulgar, pasta and they have great deals on grapeseed oil. If I wanted to actually eat – I would go to Ararat over 88 Market any day.
I think the whole point is that you *could* shop at Ararat or walk your kid to the Free School, and not so much measuring whether you *would* or not. It’s not a measure of preference, it’s a measure of utility, proximity and accessibility.
But, it would be nice if they eventually folded in some sort of qualitative measurement of the actual utility of a place beyond simple categorization, of course predicated on the people rating the utility would actually need to check the place out first.