Look into Open Elevation as a replacement for Mapzen point service

[jhollist]

https://github.com/Jorl17/open-elevation/blob/master/docs/api.md

[mpadge]

hey @jhollist, i only just discovered this great package. In the meantime, I've incorporated an elevation data function to append to data from Open Street Map via my osmdata package. I originally intended to use Open Elevation (hence me commenting here in this issue), but that repo was almost abandoned, and there was a warning not to expect the server to function. It is evidently back up in the meantime, but it's still very slow, and highly restricted in terms of data volume - a POST request for more than a few hundred points becomes almost unmanageable. Plus, it uses (at least used to use) the old 1km-res GGIAR data.

All that led me to replicating the same functionality locally using a pre-downloaded CGIAR tile (at 90m res). My code is here - feel free to crib as you like, or I could PR here if you like. Or you could just do it yourself coz it's dead easy. The disadvantage of having to pre-download is massively outweighed by the advantage of unlimited access to point-based elevation estimates. I've also got an extra function to get the CGIAR tile index. When i developed the code it was not possible to do any scripted downloads, but things might have changed, and it may even be possible to automate the whole thing now?

[jhollist]

@mpadge thanks for the ideas! I definitely want to expand the sources for this, especially the point data. The download doesn't bother me that much either. Having a global, easily accessible point elevation service outweighs any inconvenience.

I am hoping to spends some time in the coming months thinking about elevatr again and will take a look at this as an option. I am not oppose to a PR, but would want to think about how best to have the new functionality incorporated into the current code.

Thanks again!

[mpadge]

no worries - ping me any time, and all cool about (non-)PR, especially since i haven't even looked much into your code, and don't imagine finding much time any time soon, so very happy to wait and see what comes of this. One thing i tried was raster interpolation, but i couldn't figure out a way to make that sufficiently efficient, so that could definitely be something I'd love to help with here if you decide to pursue that.