KU versus KSU – the OpenStreetMap showdown


Kate Chapman asked on the OpenStreetMap US Facebook page where all we would be mapping this Labor Day weekend. I’m in Lawrence, KS right now so that’s where I’m mapping. I was surprised to see that the KU Campus is not very well represented in OpenStreetMap at all — unlike most university campuses I’ve seen. I was lucky enough to get here just in time for the start of the college football season to at least put in the Memorial Stadium and surrounding lots, but there’s lots more to do. As fellow OpenStreetMapper Toby Murray pointed out, neighboring KSU is much better mapped. I’m sure that I would feel more compelled to improve the situation would I live here, but it does make me stop and think about how a city with a major university apparently does not have many active mappers.

I am always curious about local OpenStreetMap dynamics, so I decided to do a visual and numerical comparison between these two Kansas university towns. I made two visual comparisons using a 0.01 degree grid. For each grid cell, I calculated the mean version and the mean age of the way features — age being the time that passed between the last edit and Sept 1st. On top of the grid, I projected the way features themselves in thicker, transparent lines to create a sense of feature density. Here’s the results:

Mean version:

Mean age:

The KSU campus, with much detail in the buildings and footways.

A visual inspection of these images shows that Manhattan has received more recent OpenStreetMap contributions overall. Lawrence seems to have a little more green areas in terms of mean version, although not in the parts of town that really matter most: downtown and the campus area. In terms of way feature density, Manhattan shows much more density especially around campus. This really shows when you zoom in on the campus, where a lot of detail in the building outlines and footways shows.

This visual analysis is nice to get an initial impression, but you really also need a numerical analysis to back it up. I did a quick analysis run collecting some stats for the data for both towns. It turns out that Lawrence has 0.4% non-TIGER ways, whereas Manhattan has 0.77%. These are ways that do not have any TIGER source tags. That most likely indicates a way that was not part of the TIGER import, and was thus added by a human OpenStreetMap contributor. Of the ways that have TIGER tags, Lawrence has only an average 1.31 increase in version, while Manhattan shows an average 2.56 increase in version. This means that the TIGER imported ways in Manhattan have received more subsequent attention by human contributors than in Lawrence. Also, of the TIGER ways in Manhattan, only 8% are still in their untouched 2007 state. In Lawrence, that figure is 19%.

This is just a quick analysis and as such does not answer the real question of why Lawrence’s OpenStreetMap data is so much worse than Manhattan’s? There’s a lot of potential answers to this. They all boil down to an unbalance in local contributors: Lawrence is a sizeable town with a 30,000 student university, and it seems none of those students are actively mapping the KU campus — or any other local highlights for that matter: I was at two local cafes today that seemed popular with the student crowd and neither of them were in OpenStreetMap. Manhattan, 80 miles to the east, has a somewhat smaller university but is much better mapped.

Local OpenStreetMap quality still varies a lot. It only takes a handful of contributors to improve the local data, but if there’s nobody who generates awareness locally, the momentum will not be there. So someone here in Lawrence, in the KU Geography department for example, get on the task and bring OpenStreetMap to life here! There’s only so much I can map…

4 thoughts on “KU versus KSU – the OpenStreetMap showdown

  1. Wow this is great! Never thought a little inter-collegiate ribbing would lead to such an interesting analysis🙂

    I notice that a lot of the green areas in the mean version image of Lawrence are along the river and interstate where I along with remote mappers (read: NE2 working on interstates) have done a fair amount of work. Also, the mean version in the middle of Manhattan is lower than in some of the surrounding areas because of all the v1 objects on the KSU campus that were traced off of high resolution aerial imagery from the county GIS office. (thanks for that!)

    The 0.77% non-TIGER ways kind of surprised me. I figured it would be higher. But I guess there were a lot of TIGER ways to begin with, plus you are including some of the land outside of the city where not much work has been done.

    Thanks for this work!

    • I think it would be interesting to do Intro to OSM talks at rival universities in the same week. Then watch the mapper showdown. A&M vs. Tech, NC State vs. Duke, Ohio State vs. Michigan; these are all natural interest-builders.

  2. Ahɑa, its fastidious conversation about this piece of writing here at this website, I have read аll that, so
    now mе also commenting here.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s