3 ms·
About 10 years ago I helped build a piece of software that would use OSM data and that data needed to be precise. Turned out that a lot of the Tiger Maps scans
by IgorPartola 5y ago
About 10 years ago I helped build a piece of software that would use OSM data and that data needed to be precise. Turned out that a lot of the Tiger Maps scans were not super accurate so I urged our company to hire a data entry team to clean up a bunch of it. We had I think 5 or 6 people whose job it was to precisely clean up OSM data based on satellite imagery in a bunch of small-medium cities as well as Chicago. Everyone benefitted from this which was awesome.
- tomrod 5y agoIn today's AI/ML saturated landscape, your approach to data labeling to augment applications is much needed. Awesome to hear it went well!
- tverbeure 5y agoThe TIGER data was (is?) terrible in terms of exact coordinates, but the graph of the streets is usually accurate. Back in 2007 or so, I spent hundreds of hours aligning TIGER data to the satellite imagery in OSM, usually in cities I never had been before. If I remember correctly, I did most of Fresno. :-) I’m totally fine with my edits being used commercially too.
- maxerickson 5y agoThe status of the initial Tiger data varies by county. Many counties were poorly aligned, but many counties were also pretty good, depending on whether the county had been updated in a quality improvement project that Census was doing when the OSM import happened.