Should you contribute open data to OpenStreetMap for free?

This is a guest post by Cj Malone, Director at OpenStreetMap UK. Opinion are his own and not those of OpenStreetMap UK. Cj is an avid OpenStreetMap (OSM) contributor, mainly in the United Kingdom (UK), but sometimes fixes broken data in other regions.

Cj’s sent me an impromptu response to my article Should you fix errors and contribute to Google Maps for free? In my article, I discuss the pros and cons of contributing my unpaid labour and data to Google Maps. Cj shares some interesting perspectives on the topic.

“Should you fix errors and contribute to Google Maps for free?”

I feel the same way about OSM. It may be mainly made by individuals in there spare time, but the big tech companies are making millions of it. The same as free and open source software (FOSS).

Apple uses OSM data in parts of the world where their commercial partners don’t provide data. They are reasonably good at fixing and contributing data to OSM in those regions. They employ nobody in the OSM community.

Microsoft Bing uses OSM in several regions, and is slowly moving from it’s traditional providers to OSM globally. They provide machine-learning (ML) datasets that they have computed from areal imagery. They employ nobody in the OSM community.

Facebook uses OSM world wide. They do a lot of quality-assurance (QA) work on OSM data as they were burned by a malicious user changing the name of New York City to Jewtroplis. As part of their work they now release a dataset called Daylight. Daylight is basically OSM data (+ other Open Data, like the Bing buildings) delayed with QA tests. They employ nobody in the OSM community.

Amazon do a lot of work globally to help their delivery drivers, mainly by mapping new residential roads and driveways. They employ nobody in the OSM community.

Ordnance Survey, the National mapping agency in Great Britain (GB), now uses OSM for all data outside of GB, and some data inside. Comically, some of the OSM data they use inside GB is because they couldn’t agree on a licensing agreement with themselves. They employ nobody in the OSM community.

— and so on. You can find more details and examples in the OpenStreetMap wiki.

“Google has since sent me monthly unsolicited emails about the change where they say how many people have viewed the updated listing. At the time of publication, Google says I’ve helped 90 000 people find the eye clinic in the last six months.”

I’m not gonna lie, I kind of like that idea. In the OSM world you just have to have faith that your edits are helping people. There is basically no feedback loop.

“My edit helps Google to know more accurately where (and thus why) its customers visit the hospital. My edit is actively harmful to other patients’ privacy.”

I think Google gets more statistics based on people destinations from navigation snd clicks on the POI than raw GPS. But I don’t use Google Maps, so I’m not sure.

“Then again, not fixing the location marker could cause delays at the clinic and actively harm people’s sight and long-term health prospects.”

This is a great example of why service providers should ensure data about them is up-to-date in third party places. This obviously goes back to the monopoly advantage, but if a service provider made their data available openly it could be used by everyone.

“I’ve since had a look at OpenStreetMap’s editing interface, and let me just say that “ease of use” wasn’t a design requirement.”

I’m assuming you’re talking about the iD editor on the project website. But honestly, that could be about any of the editors. The only one I can recommend as being easy to use is StreetComplete for Android, but you can only add data about existing OSM places (like adding opening hours to a shop, but not adding the shop itself).

“As an advocate of open source and open data, I should have contributed the missing information to the OpenStreetMap project. There’s only so much time in the day, though. Google only asked for a few seconds to update its map. It was quick and easy.”

Yeah, I don’t blame you, OSM has a massive learning curve. And the overlap between Open Source and Open Data is surprisingly small.

“However, their real strength lies in having convinced businesses and customers early on to contribute detailed information about everywhere.”

I’ve been working on a Places dataset by scraping Linked Data and Microdata from first party websites. But due to issues with the source data going off spec means I basically have to write some code for each site anyway. Plus, I don’t have the resources to do a large scale web crawl to find independent shop websites.

And even then, this data is probably covered by Database rights in the UK and European Union so it can’t be used in OSM without the data owners permission. It’s surprisingly hard to get companies to give you permission to copy data from there website, it’s basically free advertising for them, but they are still reluctant.

In the UK we’ve reached a mid-point, we use the dataset for QA, and to identify areas that need an independent survey to collect missing data. Depending on the interpretation of Database rights, even that may be infringing.