What does 2 hours of work give us?
What does 10 hours of work give us?
Query WikiData, limit to geoarea. Color code red-no image, green already has an image.
Link to how to get to places? Trafiklabs can do that for us. Limit to places that are no more than x minutes away and back?
In one hour I'll be able to create a quick and dirty map with a webserver deoloyed on my own server farm at kodapan.se. The backend would use a Lucene search engine filled with the Naturvårdsregistret data rather than Wikidata. Markers in the map would display a title on hover (or at all times) and link the the Wikidata entry on click.
To have the backend gather data from Wikidata and refresh the information in the search engine every now and then could take perhaps two more hours.
To request server space on WMFlabs or something and set it up there would take ... Well you probably know that better than me. Two more hours?
To execute using Wikidata as live backend rather than an ad hoc search engine will be much slower for the user and add weight on Wikidata, but would then just be a static webpage with no requirements to build a backend that needs to be hosted somewhere. It would probably take me at least two hours to implement. Mainly because I don't know how to place a SPARQL-query that gathers all items within a coordinate envelope.
With five more hours I can send a request to https://www.trafiklab.se/api/resrobot-reseplanerare and ask for a "golden API key" and implement a feature that keeps track of what items are fairly reachable using public transportation tells user how to get there and back from their current location.
The last comment is rather exciting longterm (is it open data?) but as several transportation authorities in Sweden recommends avoiding public transport (at least SLL in Stockholm) I would propose not doing that at this point.
Having it hosted on WMFLabs could be of value, especially if we want to "internationalize" it.
I don't exactly know what live backend means. What is pros and cons of that?
The backend is where we store and select the data to be sent to the user. We could just send all 42MB of data from Naturvårdsverket to the user and have them process it locally on their own hardware, but it would be terribly slow for users with old machines or for those that are using their phone. Instead we want to keep this in a datastore of some sort on a server that with speed can limit to the items available in the geographic area the user is currently looking at on the map. That is the backend. I can implement an ad hoc solution using Apache Lucene in minutes. It would be super fast. A few milliseconds to get a response visible in the map. The second suggestion of mine is that we could query Wikidata to get the data. I suppose it will take a couple of seconds to get the data back. Everytime a user moves the map, zooms, etc, we need to place a new request for data. So I suggest we go with our own ad hoc solution for speed.
Rather than trying to build a reusable configurable framework for the future, I would suggest just making it quick and dirty to fit our need here and now. If other chapters want to use it, let them come back to us and we'll adapt it for them in an hour. My estimate is that a well implemented and documented reusable solution that is easy to configure and all that would take several days to implement. So perhaps that should be considered first if the end users considers it to be a successful tool.
Here is the one hour map, zoomed in on Stockholm.
https://wle2020.kodapan.se/#59.30908,18.12444,14z
Due to a hard disk crash at openstreetmap.se yesterday, the actual background map is served from my personal maps server. It can be slow if you are the first person to look at a part of the map that nobody requested before.
There is no color coding, no links to Wikidata, etc. Just the data from Naturvårdsregistret.
If you need to stay at zoom 10 or more to get objects. It still loads lots of data to your browser at zoom 10, try and you'll notice how it kills your computer with thousands of polygons and markers in the map. It's possible to improve this a lot, e.g. by simplifying polygons on low zooms and what not.
Biosphere areas are sort of killing the features of the map too. Here's Blekinge arkipelag:
https://wle2020.kodapan.se/#56.0956,15.69122,11z
What are we asking people to photograph here? In any case, biosphere areas should be the first thing painted on the canvas, so that clicking on other polygons within the area is possible.
Maybe related task see T234196: Kan Wikipedia användas för att beskriva svenska skärgården
I spoke with Fredrik by accident and sent over some information and I feel Wikidata/Wikipedia is a perfect match... no action done since Sep 30 2019 but it looks like they got some money "Arc Gate ska öka servicen i småhamnar och skärgård"