Sun 13 Sep 2020

Q4 quarterly project 2020 ideas

Gareth L Public Seen by 50

It's that time again. The Q3 project of cycle infrastructure will end in early October. Please share your suggestions and ideas for the Q4 project.

Good ideas often include those with an underutilised, suitably licensed, data set.

Ideas have been listed in the wiki page for this, but do not feel limited to it. https://wiki.openstreetmap.org/wiki/Talk:UK_Quarterly_Project


lakedistrict Sun 13 Sep 2020

  1. EV charging points - has been suggested before and already on the ideas list.

  2. Heritage features such as listed buildings and heritage plaques?


Cj Malone Wed 16 Sep 2020

Defibs could be a good one, they have a clear benefit to communities, and have the potential to get some attention from outside the OSM community. Robert Whittaker already has some tooling for surveying them, and we could create a tool to make adding them easier for none OSM experts.


Cj Malone Mon 21 Sep 2020

Guernsey have a map of there defibs overlaid on an OSM base map, I'm gonna email them in the morning asking if we can work together to get them into OSM.


Robert Whittaker Fri 18 Sep 2020

I've recently been having a play with the Retail Points dataset from Geolytix. (Now available under a licence equivalent to the OGL.) It provides locations and addresses for supermarket and other food outlet chains in the UK. No maps yet, but iv'e got some tables of OSM comparisons and discrepancies up at https://osm.mathmos.net/chains/ . "Brands" might make a good QP. We could work on: getting more of the Retail Points chains mapped (a missing branch of a chain often indicates other things nearby need mapping), getting more brand:wikidata and store website tags added, generally improving the consistency of how branches are mapped, and also on improving the Name Suggestion Index (the think that powers the iD brand suggestions). With a bit of momentum, we might also be able to use the project to help persuade other chains to release their branch data to us to help further improve the map.


Cj Malone Fri 18 Sep 2020

An issue with OGL is that there may be other copyright on the data, and if so it overrides the OGLv3. Does the equivalent license from Geolytix also have that issue? I've done similar things using data scraped from individual store finders when I was trying to update Retail chains in the United Kingdom, but I was careful to just improve the existing OSM tagging (eg adding brand:wikidata) or adding stores I could see via Mapillary, not using the data as a source to add new things.

But I agree that this is a good one, and the arm chair aspect of cleaning up brand tags could be perfect through the winter.


RobJN Mon 21 Sep 2020

I saw quite a bit of back and forth between Robert and Geolytix on Twitter for this so hopefully it is now all good from a licence point of view. Hopefully Robert can shed some light on the situation here.


Dennis Bauszus Thu 24 Sep 2020

That's correct. Robert gave us (I am working for GEOLYTIX) some pointers for the licence to amend which we have done. The data should be good to go into OSM.


Gregory Williams Tue 22 Sep 2020

I've found myself adding quite a few playgrounds recently. On the ground the different types of playground equipment can be added. It's also usually fairly easy to spot a playground in aerial imagery. When adding playgrounds from aerial imagery I've often spotted other issues, such as missing paths, or spiky roads where their shape could be improved.


Dan S Tue 22 Sep 2020

I'd like to argue against playgrounds as a quarterly project. Just like zoos and theme parks, it seems people are happy to map them in exquisite detail, under their own steam. I don't think we need any co-ordinated push on them.


Dan S Tue 22 Sep 2020

Stephen Colebourne discussed on talk-gb tagging modal filters. If we can get tagging agreed, I think that would be a great, useful and timely thing to do.


Huw Diprose Wed 23 Sep 2020

I suspect this is one that must have been tried before, so i'm interested if there's history.

But it seems we have a gap around political boundaries. For instance I've been trying to query overpass turbo for nodes within my ward. But it appears that data is only partially present across the UK

I'd suspect this data should be easy to obtain and license.

I think it's important because it would allow users to connect ground truth features in an area to political representatives that might be able to change them. Perhaps a lack of amenities in one ward could be connected to a representative who could then be asked to take action within local government?

I'm particularly interested in wards, but would suggest a candidate list could be something like this:

  • Local government 'electoral division' (wards for parish / district councils, electoral division for county councils) boundaries

  • Directly elected mayors - Boundaries here are likely to be more complicated, in some cases I suspect it maps to the administrative borders of a local authoirty, in other cases it may unite a few (eg London boroughs?)

  • Perhaps police and crime commisioners for Englad and Wales (they map to the 41 police force areas, not sure if there are already administrative boundaries?)

  • Devolved government constituency boundaries (NI Stormont, Senedd, Scottish Parliament?)

  • Parlimentary constituencies - Not sure if these are already mapped?

After that I somewhat lose track of other forms of democracy, city of london has some unusual mechanisms, and there's some older stuff like aldermen and honorific roles i'm even less sure about.

Apologies if i'm re-tredding preivously considered territory here. Happy to take a previous discsusion to go read up on?


Dennis Bauszus Thu 24 Sep 2020

I agree that it would be nice to have political boundaries in OSM. It is a huge pain though to continuously maintain the validity of those boundaries and easy to break boundaries while editing features which are tied to the boundary. I have myself fixed the French and Spanish constitutional admin districts a few years ago and it is a very manual labour. Happy to participate looking at UK admin boundaries.


Gareth L Mon 28 Sep 2020

I'm going to suggest waterways/courses.

I've been doing some mapping in Wales (family nearby) and it highlights to me how much stuff traced from NPE is still there. Waterways change a lot.

There's opendata streetview layer which is very useful for this. The latest Bing imagery is, on the whole, very well aligned when compared to multiple gps traces and streetview layer.

The clarity available is allowing many bridges/culverts and redirected waterways be picked up.

Progress on this might be tricky to quantify though.


Gareth L started a dot vote Mon 5 Oct 2020

Which suggestion(s) could make a successful Q4 project? Closed Thu 8 Oct 2020

by Gareth L Thu 8 Oct 2020

By a clear margin, AED/Defibrillators have been selected. Thank you for voting!

Considering the discussion above, and with the resources available.

5 - EV charging points
8 - Heritage features/buildings/plaques
13 - AED/Defibrillators
8 - Brands/Chains
7 - Modal Filters for ways
3 - Political/Administrative boundaries
9 - Waterways
7 - Playgrounds equipment

RobJN Sat 10 Oct 2020

Useful resource for anyone wanting to see if we can get hold of more data for the quarterly project. Let us know if you intend to contact any of these. https://www.suddencardiacarrestuk.org/uk-aed-map/