osmlatvija / osmalyzer Goto Github PK
View Code? Open in Web Editor NEWParsing OSM data in Latvia against various data sources
Home Page: https://osmlatvija.github.io/Osmalyzer/
Parsing OSM data in Latvia against various data sources
Home Page: https://osmlatvija.github.io/Osmalyzer/
I am parsing this from website maps, since all banks prominently mark this, but I'm not using to check if OSM elements have the correct tag(s) set.
Explain the icons found on the map in a legend, especially correlator stuff.
Stretch goal would be to allow the legend to toggle issue layers, although not sure how that would work with clusters.
Instead of having a hard-coded file/list - NSI index
Let user manually mark a problem as resolved when they know it's not actually a problem with OSM, but with data or data interpretation. This way confirmed "problems" don't report as problems on the map and clutter the view.
For storage, current plan is something like Google Spreadsheets.
UI will have to be JavaScript (github.io is static pages, so cannot do anything in backend).
Basically, all the tags and whatnot plus links to common stuff. Currently, you have to open the link to see. This will really bloat the report page though, possibly needing a JS call. Or filter to relevant tags - but then it's a lot of manual hard-coding work.
Currently, public transport issues are really hard to parse and understand. The problem is that there are so many issues with so many routes and it's all on the same map, but then each issue type is its own map. It's all really complicated though, so not sure what the best approach would be. Perhaps map per route with all issues combined for that route? It needs to get converted to correlator too.
Stuff in data\ folder.
Treat like analysis data, i.e. local data?
ref
Full government list at https://www.zva.gov.lv/aptieku-karte/ although the site is dead at this time - last archived Jan 2022
Open data at https://data.gov.lv/dati/lv/dataset/farmaceitiskas-darbibas-uznemumu-registrs , although it's hosted at https://dati.zva.gov.lv so it's still dead right now...
May be full data is more usable without needing individual lists, although I still probably want per-brand reports or it might be too many. Not sure if it includes irrelevant locations.
Besides regular links, https://relatify.monicz.dev would be useful for missing stops, broken connections, ambiguity and such - really nice visualization and editing. Of course, those issue checks still need to be implemented.
Latvijas Pasts mail boxes have this validation tool https://mapillaryby.github.io/PostboxValidator/latvia_index.html
LP Express in the future planning to install lockers.
Find ways that don't interconnect because they terminate at the edge of an area - parking lot, pedestrian area, etc. This is not really Latvia-specific though. There probably are QA tools for this already. No sure if it's worth it.
Shop websites generally list their individual shop opening hours, which presumably are more or less up to date.
Would need to parse them - I did not parse any when doing #3.
https://www.depozitapunkts.lv/
Problem is not all are actual separate POIs - a rural shop may accept them manually.
Currently, all reports are in English. I imagine they could have 2 language version - Latvian and English. Explaining things, especially to new editors, would be clearer. This is a lot of extra work though, given that I haven't even written most stuff in English. I don't think automated tools would help, since the primary reason to explain is accuracy for fairly technical stuff.
The icon used for multiple issues at the same location is the icon for the first issue and not the combined icon variant - probably need to reuse clustered icons. Would also make sense to just show them as "clustered" if actually just one marker.
I wonder if I can clustered at lower zooms only cluster issues on top of each other - then I can just use that logic. The problem is how do I see popups for the clustered marker like that.
There are a bunch of related ones like shops and banks and transport that can be shown in subgroups.
There is a list at Valsts izglītības informācijas sistēma (VIIS) https://www.viis.gov.lv/registri/iestades
Searching without criteria brings up the full list, paged.
They provide a CSV export, but it doesn't have coordinates/location or IDs.
I can get the "searched" list as JSON from https://www.viis.gov.lv/registri/iestades?search=true&json=true, which gives me nearest page links (basically for UI), which then provide something like https://www.viis.gov.lv/registri/iestades?search=true&json=true&page=3 that I can keep iterating with full data on each page in chunks of 20, which means 110 pages at this time.
I can also get an individual record JSON from an ID at https://www.viis.gov.lv/registri/iestades/25. This is what the website loads when clicking an entry.
Some details:
data._rows
has actual "full" data
data._rows.total
for count
data._rows.hits
are the "search" results
data._rows.hits[0]._source
is then the data of each record
notable immediate fields are Name
and InstitutionTypeName
each also has data._rows.hits[0]._source.locations
notably these have LAddressType
and LStreet
and some others but no actual street number, like it will have "LĒDURGAS IELA" and "JŪRMALA" but no number.
I have to get the individual record, which then has data.pamatdati.papildinfo_telpas[0].AddressFullText
with usable address. But this means I have to get each record for all 2182 records as of now... which is not feasible.
Match data to OSM elements based on how well they match and report accordingly. Two main reasons.
UI can show more matches without more false positives. For example, a data item with an address could match an ATM with that address exactly even if it's like 500 meters away. But normally, it would only match that ATM without further data close nearby.
Another example is when a better matching element is further away than some closer but weaker matching element. For example, a cultural monument ID would be a perfect match, while having operator only a good match and only being tagged with heritage a poor match. It should then pick a stronger match rather than closer match.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.