The Washington, DC region is great >> and it can be greater.


How school tiers match up with Walk Score

One of the best effects of open data is when people correlate data sets from very different places to generate interesting information. This graph cleverly combines DC's school quality tiers (known as "accountability categories") with Walk Score:

Sandra Moscoso wrote yesterday about how Code for DC's School Decisions Project has been gathering coders who want to use open data to help parents, students, and policymakers. This is one of the graphs they created at the recent Open Data Day using data from the Office of State Superintendent of Eduaction (OSSE).

I've asked to get access to the raw spreadsheet for this graph so we can look at, for example, which schools each dot represents. Here are the accountability categories by school. I will add the spreadsheet with WalkScore matched up with category when it's available. Update: here's the data as a CSV file.

A few things immediately jump out. The most successful DCPS schools have high Walk Scores, while the least successful ones mostly (but not entirely) cluster in the lower range. This may reflect the fact that a public school's success has a lot to do with the socioeconomic status of the neighborhood, and the local retail that is a big part of Walk Score locates in areas with higher incomes.

That income effect is also very pronounced in the graph Sandra posted yesterday:

That's not the case with charter schools. 3 of the 5 "reward" charters are in low-Walk Score areas (which could mean something, or just be a consequence of little data), while the "Rising" charters are basically all over the place. This may have a lot to do with the simple fact that since charters have to find and pay for their own space, they're in all manner of locations.

An interesting future step might be to correlate the school tiers with some data set about land prices or rents, or resident incomes. That could help illuminate whether charters end up locating in less-expensive areas, because they want to serve poorer residents and/or because they need cheaper land.

What do you see from looking at this data?

David Alpert is the founder of Greater Greater Washington and its board president. He worked as a Product Manager for Google for six years and has lived in the Boston, San Francisco, and New York metro areas in addition to Washington, DC. He now lives with his wife and two children in Dupont Circle. 


Add a comment »

schools are a part of the walkscore, so where schools are clustered within X miles of each other, there's gonna be a score distortion.

Also, if the hope is to actually demonstrate a correlation betw walkability and performance, hope that walkscore's "street smart" beta version is used, which actually measures network distance v. crow's flight

by darren on Mar 13, 2013 3:40 pm • linkreport

The first thing I see is that I have no idea which are the good schools and which are the bad ones based on the names for these categories. Reminds me of the Big Ten naming its divisions "Legends" and "Leaders" instead of something common sense like "East" and "West."

by Geoffrey Hatchard on Mar 13, 2013 4:16 pm • linkreport

@Geoffrey Hatchard -- I suppose that's the problem with having to use the labels you're given.

In terms of test scores, the lowest-performing are on the left and the highest-performing on the right.

As for the B1G, I have no answers on who is a leader, who is a legend, or where Rutgers fits on that spectrum.

by Jacques on Mar 13, 2013 5:08 pm • linkreport

@Geoffrey Hatchard: While I agree that the labels aren't very clear, I think it can be easily inferred from the map. Purple dots in the west (of Rock Creek) can't meant low performing

by Simplicity on Mar 13, 2013 6:31 pm • linkreport

I see something to practice a logistic regression on. Looking forward to the spreadsheet.

On the charter schools side, you have too few observations to be meaningful in the highest and lowest quality tiers. In addition, don't charter schools choose where to start up? They aren't scattered randomly throughout the city. If there is no correlation between walk score and quality for charter schools, then that is interesting. I would wonder why. From the DC map of charters, the charters have started in every section of DC apart from the richer parts of Northwest DC (which are the highest income, aren't they?).

by Weiwen on Mar 13, 2013 6:57 pm • linkreport

B1G is likey going to change the division names for 2014.

by selxic on Mar 13, 2013 9:51 pm • linkreport

Thanks for posting this, David! I'll send you spreadsheets this afternoon (out and about this morning).

It's worth pointing out that the accountability categories are fairly complex, and based on both absolute and trends in test scores, as well as the presence or absence of sub-groups (demographic) that score particularly poorly. So yes, the categories are ordered in some sense, but don't think of them as quantiles (lowest 20%, next 20%, etc.).

And no, we didn't use the new Street Smart walkability scores. It takes a while to get an API key from Walkscore, so we actually just did hand data-entry to pull scores from school street addresses for this. In the future we'll be sure to check out Street Smart, though!

by Harlan on Mar 14, 2013 7:47 am • linkreport

because the charter schools are not neighborhood dependent for students, of course they won't show much correlation on walk score factors in this kind of study.

Plus, charter school location decisions are for the most part "driven" by property availability.

Which again has little if anything to do with spatial pattern except inversely--industrially zoned land is most likely to have available properties and/or charter schools can outbid economically-decision-based industrial property owners for the property--and the land is likely to be in locations that are less walkable, because it's industrial.

Of course, this is less the case when DCPS deaccessioned schools are used by charters. But my sense is that these schools are more likely to be located in less walkable areas (because high income households are less likely to seek out such locations) and because of less density.

by Richard Layman on Mar 14, 2013 9:45 am • linkreport

Richard, thanks, that makes sense! I suspect with longitudinal data one could actually make a strong causal-statistical argument, but I'd be surprised if the property-cost issue wasn't a substantial part of the reasons for the pattern!

by Harlan on Mar 14, 2013 3:00 pm • linkreport

Guessing at the actual coefficients here, it strikes me simply that public schools, which many students attend simply because they're nearby, reflect the character of the neighborhood. More affluent/educated people are more likely to locate in more walkable areas (I believe that's pretty settled, except for VERY rich communities that aren't particularly walkable - but many of those kids are attending private schools and/or have household staff or a parent at home to drive them to school), their kids are better prepared for school and better supported in their educational pursuits, and, hence, the schools score higher. Because the schools score higher by virtue of who lives in the neighborhood, out-of-bounds students are similarly motivated/prepared/pushed by virtue of the tough competition to get in. But, as has been noted, charters go kind of wherever, and people put their kids in those charters without regard to their location, when the charter meets their needs.

Overall, I don't think that walkability makes schools better. Sure, it makes them more convenient and safer to get to, which are BIG benefits, but it seems a stretch to imply that walkability alone makes schools perform better. I know that wasn't implied here, but for anyone who was thinking it, it strikes me that the causation, if there is any, goes the OTHER way.

by Ms. D on Mar 14, 2013 10:00 pm • linkreport

Add a Comment

Name: (will be displayed on the comments page)

Email: (must be your real address, but will be kept private)

URL: (optional, will be displayed)

You can use some HTML, like <blockquote>quoting another comment</blockquote>, <i>italics</i>, and <a href="http://url_here">hyperlinks</a>. More here.

Your comment:

By submitting a comment, you agree to abide by our comment policy.
Notify me of followup comments via email. (You can also subscribe without commenting.)
Save my name and email address on this computer so I don't have to enter it next time, and so I don't have to answer the anti-spam map challenge question in the future.


Support Us