RMK (State Forest Management) recreational points of interest + data liberated guerilla style

[tambeta]

The points of interest of RMK (State Forest Management Centre) should ideally be open and machine-readable, but as a part of a project I have written a scraper to pull an occasionally updated JSON dump from the web.

[tormi]

Nice! BTW, GitHub can show your json as a map if you were using geojson format instead. See this for example: https://github.com/buildig/rkvr/blob/master/hooned4326.geojson

[andreskytt]

How stable is the dump? What would you need to make it not-occasionally-updated?

[Hillehinsberg]

@tambeta could you provide some user cases? - what would you do with this data when machine-readable?

[tambeta]

@andreskytt Basically, the dump gets updated whenever I start wondering if any new hiking paths have sprung up, which might be a couple of times per year. I'm not sure this dataset should be maintained and regularly updated, I'd say it makes more sense to pressure RMK into opening up their data.

@Hillehinsberg Well the map application I built on top of this scraped data is a good example. The official site doesn't have such functionality, or at least did not have when I started this project.

[tambeta]

Updated now.

[andreskytt]

How often does RMK change their web? We'll pressure them for sure but there's hosting space available with some cron robot if necessary

[tambeta]

In practice, I do update the dump, but wouldn't like to attach any guarantees to it. This doesn't need to happen very often, because the points of interest are by nature quite static. The problem with cron jobs is that the crawling result does need some manual curation, because sometimes coordinates are wrong, the page structure has changed, some assumed invariants do not hold (e.g. "Telkimisala" is suddenly called "Telkla" on some random POI pages) etc.

[tormi]

In addition to POI-s there are also hiking tracks (= lines in terms of geometry). Also, "Telkimisala" can be point or (more likely) a polygon. So, RMK-s action is needed here.

It would be ideal if RMK could open up it's data about POI's and tracks in a GeoJSON format. It would be even more ideal, if they'd use an OpenStreetMap. There is already a lot of crowd-sourced data available at OSM, for example http://www.openstreetmap.org/relation/6354083.

By using http://overpass-turbo.eu/ we can query and download all the data where operator=RMK:

Query:

relation
  [operator=RMK]
  ({{bbox}});
(._;>;);
out;

Output:

• live map
• screenshot
• data exported as GeoJSON to gist.github.com

[tambeta]

Actually, my rmkmapper.pl script has the ability to parse tracks as well, and convert these into JSON (~albeit not GeoJSON yet~ everything is now output as GeoJSON), with the help of GPSBabel. This works well with the GPS tracks provided by RMK.

[tormi]

Good to know RMK provides tracks data for 2 major routes: https://loodusegakoos.ee/where-to-go/hiking-route. But does it provide gpx files for all the other tracks it maintains?

[tambeta]

@tormi Not that I know of. But nowadays they often have PDF maps of their tracks so some volunteer effort could produce machine-readable tracks from these. If some geography teachers were on board with this, it would be a good practical task for hands-on geography lessons in schools. Doesn't require much more than Google Earth, software-wise.

[tambeta]

I worked a bit on this project and it now outputs GeoJSON for both POIs and tracks. In addition the git repo contains fresh dumps of both (note that tracks are just the two long "matkatee" trails).

[infokujur]

Need to bring this knowledge to RMK ppl.

[infokujur]

Julgustan rakendust üles laadima @tambeta avaandmete portaali. Avasin andmete enda jaoks uue issue https://github.com/okestonia/opendata-issue-tracker/issues/153 , mille suuname RMK inimestele.