## [Python] Geographic heights are incredibly easy!

From: andrew cooke <andrew@...>

Date: Sun, 13 Jan 2019 20:47:20 -0300

Wanted to add heights to bike rides in Choochoo, given the GPS
coordinates.  At first I considered stealing the data from Strava, but
their T&C don't like it and anyway I couldn't find it in the API.
Then I considered Google's Elevation Service, but it's \$5 for 1,000
values.  Then I considered the free OpenStreetMap equivalent, but that
seemed to be broken.  Then I realised it's pretty damn easy to do by
hand!

Turns out that the Space Shuttle scanned the entire Earth (except the
Poles) at a resolution of one arcsecond (about 30m on the equator) and
the data have been publicly released.

The project is called SRTM, and the 30m data are "v3" or SRTM3.  More
info at https://wiki.openstreetmap.org/wiki/SRTM and

http://dwtkns.com/srtm30m/ and there's an excellent (although buggy)
blog post on how to parse the data at
code at https://github.com/aatishnn/srtm-python

Because the coords are in the file name there's no need for any kind
of RTree or similar index - you just take your coords and infer what
stick it in an array, and return the array value!

My own code to do this is at
https://github.com/andrewcooke/choochoo/blob/master/ch2/sortem/__init__.py
and includes bilinear interpolation (you could cut+paste that code
except for the constructor which is specific to Choochoo - just replace
the messing around w Constant with a literal directory string).

The tests are at
https://github.com/andrewcooke/choochoo/blob/master/tests/test_sortem.py
and from the contours there, which are plotted across 4 tiles, it's
pretty clear that the interpolation is seamless.

So easy!  I thought this would take a week of work and I did it all
this afternoon....

Andrew