Data is Infrastructure Too

Wed 31 January 2018 | tags: Open Source, Infrastructure, Big Data, -- (permalink)

Sadly, Mapzen started the process of folding up this past month. Their goals were always going to be a hard thing to achieve, but they did assemble an extremely capable team on their way. That most of their tech will live on beyond the organization itself will highlight to consumers of their technology why the approach Mapzen took was so valuable. I see many of the team have found good landing places, but it is sad for them to pass nonetheless. The entire Mapzen experiment shows that a company can rise, live, and die without destroying the technical base of its clientele if things are organized well. I hope more technology-dependent organizations realize the value of approaches like Mapzen's (and Hobu's too!) built on open technology foundations.

The episode highlights a related topic Brian Timoney tweeted about it the other day:

There are many companies rushing to capture LiDAR data for all of the highways and roads in the world. The data often looks like this street scene in Cedar Falls, Iowa (data courtesy of Iowa DoT -- click the image to visualize it Potree right in your browser).

Iowa Bridge Scene in LiDAR

Data like this is the raw material of every base map of every autonomous vehicle scenario. There are companies with equivalent captures like the image above for nearly every paved road in the United States. This data is the infrastructure that the transportation system will need to function in a not-very-distant future, but state and Federal DoT organizations don't have the resources to invest in this infrastructure themselves.

Government should be investing in data infrastructure like this because the privatization of it is going to allow rentiers to charge access to both the base data and the continual stream of updates they control. This business model should seem familiar to anyone who searches Google or browses their Facebook feeds -- the data is individually minimally valuable to you but massively valuable in aggregate. These streams will be flowing at vehicle manufacturers, data infrastructure companies, transportation/trucking, and so on. Those rent seekers are going be able to inflate the access cost to the data.

They're also going to try do things like indemnify themselves from any liability of use. Where does the tort chain go when my car (rightfully) crashes into a wall? I must have clicked some ToS checkbox that said I wouldn't be suing my data provider ... and on and on it will go.

Government has a role to play in this sphere, and it needs to get in the arena and start playing it quickly before the vendors set up all of their feedback loops. Otherwise, our nation's data infrastructure is not going to serve citizens in the way we would want it to.


comments powered by Disqus