Loomio
Sat 15 Apr 2017

Welcome! About "PROJECT Gamification & Visualisation"

R
RobJN Public Seen by 566

Introduction

The aim

Make adding data as fun as it originally was with roads (no no small task then!!)

How we got here

When OpenStreetMap started with a blank canvas it was easy to see the gaps. Now you have to scratch deeper until you find missing data and this missing data is generally seen to be of lower value than e.g. roads.

Some solutions

These possible solutions came out of our first in-person meeting.

1. Feedback & reward

The initial visual feedback reward was the standard map display. Seeing this go from empty to a road map was amazing. As the map has filled up the visual feedback has become less obvious.

  • How else can we provide visual feedback and/or rewards in order to put the buzz back in to mapping?

2. Making contributing fun(ner)

Surveying is an outdoors activity but data entry often relies on desktops or laptops. With mobile platforms what new options are available to us?

  • We have a separate project looking at basic apps. Best to keep to app ideas to that project.

3. Show a "crap map"

This relates to the feedback & reward idea above but instead of showing feedback when something is done the aim here is to make it more obvious what needs doing. A "crap map" if you like.

  • How can we better show what still needs doing in your local area? Could we better utilise the OSM Notes feature but make it more visual?
R

RobJN Sat 15 Apr 2017

Help is needed to answer the following questions:

  • How else can we provide visual feedback and/or rewards in order to put the buzz back in to mapping?
    • Do you have an example of a good visualisation that has encourage you to map more?
  • How can we better show what still needs doing in your local area? Could we better utilise the OSM Notes feature but make it more visual?
    • Do you use any tools to find out what needs mapping in your area?
R

RobJN Sat 15 Apr 2017

My thoughts:

Everyone has different interests and we can't possibly provide a visualisation of everyone (although we could create/point to a tool that lets people set up their own) so we need to focus on general visualisations.

Example visualisations:
- OSM Notes (but will need some clever ways to filter/group).
- Quarterly project tasks.
- Top 10 mapping themes
- Top 10 mapping locations (but how to visualise progress??)
- Known bad data (e.g. instances of Lloyds TSB)
- Known missing data (e.g. missing opening hours)
- Map showing where an OSM user contributes to one of the above (e.g. your OSM user logo on the map)
- Map showing edits from the last x days.

SK

Stefan Keller Mon 17 Apr 2017

See also our Kort Game (http://www.kort.ch/ ) relaunching soon as free iOS/Android app, as well as StreetComplete (https://wiki.openstreetmap.org/wiki/StreetComplete ) free Android App.

SR

Stuart Reynolds Mon 17 Apr 2017

This is a difficult proposition, because everyone's idea of what is important is different, and everyone's reasons for mapping are different. Largely because of how I use OSM, for example, I am interested in large POI (schools, hospitals), new housing estates, etc. and also in walking routes - footpaths and the like. Things that excite others, like postboxes or trees, just don't interest me. Opening hours, for example, is something that (rather like business names and types) would be something that is a never-ending micro-update.

The other difficult proposition is how you might quantify it, and measure progress. So "area has N shops of which only M have opening hours" is measurable and trackable, but "area only has N shops and there ought to be M" is something that requires (for example) access to trade directories or other sources of complementary data. If I had to pick, though, between the two then I would argue that the second case is more important (to me) to map than the first.

This is not to say, though, that what you are suggesting can't be or shouldn't be done, just that it might be difficult. I agree that we need to excite people to do more. My suggestion would be to start with big visible wins (the schools quarterly project was one such) and move down to finer things later on.

FWIW, if I want to see what might need to be done in my local area, I use Keep Right e.g. https://www.keepright.at/report_map.php?zoom=14&lat=51.5531615&lon=0.6521351

Regards,
Stuart

R

RobJN Mon 17 Apr 2017

@stefankeller Oh Stefan, you're reading my mind! I was just reviewing SotM 2017 talks and saw Kort relaunch in there. Definitely interested. Is there anything we can do now to help (e.g. beta testing) or is it just a case of waiting? (cc: @spiffy @adamhoyle). I've been watching Street Complete too :-D.


@stuartreynolds:

access to trade directories or other sources of complementary data

Hopefully this is something OSMUK can help with :-D

SK

Stefan Keller Mon 17 Apr 2017

We're not yet ready. But if I you're interested in testing pls. send a mail to geometalab(at)hsr.ch .

RW

Robert Whittaker Fri 28 Apr 2017

I wonder if we could have something like Keep Right but include more UK-specific errors/omissions that have been detected. e.g. the various missing things that my tools flag up, instances of BHS or Lloyds TSB objects that need checking, etc. Perhaps there could be a distributed model, whereby different tool developers can provide 'error' feeds in a common format format of places that would benefit from a survey. Anyone using such a site (or even an app based on it) would be able to choose which types of error to show.

R

RobJN Sun 30 Apr 2017

Hi @robertwhittaker. I agree - something like Keep Right but geared towards a less technical contributor (e.g. less geometry "errors" which often aren't errors, to be replaced with Lloyds TSB type errors). It would also benefit from an aggregated stats view, just like the postcode map you did for the Schools project.

The ability to have "error feeds" is great too. In a way I see this as a Maproulette crossed with Keep Right crossed with your Schools visualisation BUT geared very much towards local mapping.

With time the error feeds can also be picked up by an app (e.g. @stefankeller's Kort) so we need to find a way to add features to the website version. For me this is the aggregate visualisation and possibly some user stats (so pulling in @harrywood's user stats style work he did on the schools project).

Here are the next steps that make sense to me. These can run in parallel. Thoughts?

Scope

  • Agree a scope so that we have a set of standards that the error feeds need to meet.
  • Agree the rough aim of the website.

Error feeds

  • What attributes would we need from an API (if building one from scratch).
  • Review existing APIs (e.g. OSM Notes, Maproulette, Keep Right) to see if we missed anything.
  • Work out if we can integrate existing APIs.

Visualisation

  • Agree a map visualisation - e.g. do we stick with Postcode Areas, or move to smaller Postcode Districts, or something else (e.g. Parliamentary constituencies as they are of "similar" size/70k people).
  • Agree an approach to displaying user contribution stats.
R

RobJN Sun 30 Apr 2017

On region grouping the parliamentary consistencies make sense if the aim is to have equal population region. The ONS' MSOAs are too small (7200 people in each MSOA).

Postcode districts:
- population data from here.
- bucket size is 5000
- There are 2269 postcode districts
- average population per district is 25000, st dev is 16000

Parliamentary constituencies (GIS data):
- population data from the wikipedia page.
- bucket size is 5000
- There are 650 parliamentary constituencies
- average population is 71000, st dev is 8000.

R

RobJN started a check Sun 30 Apr 2017

Who want to join the working group? Closed Fri 12 May 2017

(Ooh, Loomio has new decision tools!). I have scoped up an idea to progress this. Who wants to join a group to get this going?

3 - Yes
1 - No
R

RobJN
Yes
Sun 30 Apr 2017

I'm in (obviously!)

RW

Robert Whittaker Tue 2 May 2017

I'm in too.

On areas for a visualusation, I don't think it is just population size (as a surrogate for number of mappable objects) that we should consider. The physical size of an area will also affect the ease in which it can be mapped.

I am not convinced that parliamentary constituencies are well enough known to be a suitable geometry. Local government districts might be preferable to them. But in general I prefer postcode boundaries, since the sizes are generally sensible and anything with an address is easily assignable to the right area.

A

Andreas Thu 4 May 2017

Hey @robjn . I am the developer of the new kort mobile app which is completely rewritten with React Native. The idea of the Kort game is to present the user missions nearby. These can be as easy as completing missing restaurant type information. For a full list check
https://github.com/kort/kort/wiki#available-missions-error-types
New mission types such as opening hours should also be available. In order to motivate the users, gamification elements such as rewards, achievements and highscore lists are presented.
The kort backend provides the client app with new missions nearby. It is also currently rewritten. As an 'error feed' the following sources are planned to be integrated:
- KeepRight
- Overpass

It should be possible to integrate your error feed, as long as the mission types are kept simple (e.g. there is no way of editing geometries).
When a certain amount of equal answers are delivered to the backend, these answers would be automatically registered on OSM.

A preview version should be available in June.

R

RobJN Thu 4 May 2017

Hi @andreas23. Thanks for getting in touch - the app sounds great.

Yeah, we'd use simple tasks only. Do you know if it will be possible to show/hide some error feeds so as to keep it simple? I look forward to June.

A

Andreas Thu 4 May 2017

If I understand you correctly, you like to have the possibility to filter missions by their error source? It is not yet defined whether or not this is possible.

R

RobJN Sat 6 May 2017

@andreas23 Yes, filtering by mission (or a group of missions) would be great.

@robertwhittaker Thanks. I will leave it to the end of the proposal period to give others a chance to state whether they want to help or not. After that let's get cracking on this :-)

Re: regions. Yes lots of different things at play here. Given that the data we have will be nodal perhaps we can have multiple visualisations??

IC

Ian Caldwell
No
Thu 11 May 2017

RW

Robert Whittaker
Yes
Thu 11 May 2017

AH(

Adam Hoyle (Director)
Yes
Thu 11 May 2017

BP

Brian Prangle Sun 10 Dec 2017

Not much seems to have happened here. Is there work progressing elsewhere? I ask because Streak, by Ilya Zverev might prove to be something we can promote more or less immediately. Maybe it would be nice to have a UK version confined to data for the UK

R

RobJN Wed 13 Dec 2017

I like this and agree with Brian in that we should adopt and adapt. No major adaption required, just the wording of (and selection of tasks). I don’t think we should overcomplicate the system by forcing for the edit to be in the UK. Although we prefer people to map in the UK this is not an exclusive – the main thing is that we inspire mappers living in the UK.

DU

[deactivated account] Sun 10 Feb 2019

Gamification of Footpaths:

Wasn't sure where to put this entry, so if moderated, please feel free to move it.
I live in Wiltshire, and when out walking, often see public footpaths that have fallen into disuse, and at risk of being lost, all together.

Using OS/ OSM data, I wonder whether lesser used footpaths could be revived, and UK (and beyond?) users encouraged to use paths, for health benefits. Essentially, I was thinking that this could be game-ified by users scoring points for using footpaths, and more points for treading a less used path. An additional benefit would be that OS/ OSM would gain updated tracks for mapping and statistical usage.

Do you know whether any one has looked at this idea?

Many thanks,

Chris.

SH

Steven Horner Sun 10 Feb 2019

Sounds like an excellent idea, I have sort of being doing this without the app gamification, either drawing expanding circles around where I live or walking every PRoW in a grid square.

I have been using StreetComplete on my phone a lot recently as well to help fill in things like road surfaces, speed limits etc which becomes addictive.

Being able to combine the 2 somehow would be fantastic.

Love the suggestion, just wish I had the skill to code an app to do it.

Sort of related I wish StreetComplete would allow me to fill in footpaths at the side of the road, sidewalks as Americans would call them. It already has the ability to tag cycle lanes, would think it would be easy to add sidewalks in the same way.

Steven

DU

[deactivated account] Mon 11 Feb 2019

Thanks, Steve.

'wish StreetComplete would allow me to fill in footpaths at the side of the road'

Strangely, I got the new StreetComplete beta last night, and it included this feature :-)

SH

Steven Horner Mon 11 Feb 2019

Me too, noticed it about 30 minutes after I sent that email. If only the PRoW footpath ability could be added as fast ;-)