Wednesday 26 June 2013

MomoLondon: Mobile, Maps & Geolocation — so much more than “Where am I?”

This Monday’s session was a little less lively than the previous session on mobile marketing. I expected the panel to violently disagree with each other on licensing, data costs and providing data to the public… but they were all far too polite!

The panel discussion covered a wide variety of subjects, fulfilling the session’s premise: geocoding, open data, data quality, visualisation, alternative methods of navigation, indoor location… It was a huge area to cover in a couple of hours.

For me, the main change in the last year or two has been the opening up of data sets. Open Street Map has gone from strength to strength and can now be linked up with live transport information and other citywide data sets.

As Jeni Tennison remarked, the most exciting area at the moment is putting all this data together in innovative ways. More than the Google Maps mashups of earlier years, we can now experience 3D immersive visualisations of cities, with real-time data pulsing through the streets and under the ground — like ViziCities, for example.

Panel

How easy is it for devs to turn lat/longs into human-readable info?

  • HW: hard work to go to postcode or nearest POIs
    • getting hold of open data is important
    • is displaying the map the best thing?
    • FourSquare did really well with a location-based app that didn’t show a map
  • GG: location powered by FourSquare has become a heavily used API
    • it’s free to access but large amounts of usage lead to business negotiations
    • similar for other geocoding APIs
  • JT: Open Street Map allows us to choose what we want to mark on the map
    • Ordnance Survey was mainly restricted to Churches and Pubs…
    • Open Street Map has shops (and pubs!)
  • IH: how do you ensure that all the data is there?
  • HW: it’s a problem that goes away with time
    • always find your nearest pub? getting there in the UK, but some patches in the north with less coverage
    • but the map gets built out on an interest-first basis, so better coverage in more interesting/populated areas

How much do we trust what we see on our mobiles?

  • GG: in US a woman sued Google ‘cos the walking maps told her to walk across all 16 lanes of a superhighway…
  • HW: we trust it too much in some cases
    • OS really shines with quality of data
  • IH: Google & Apple’s branding has gone a long way to encourage trust in map data
  • CO: Apple chose not to visualise lots of data on the map, but they got panned for it
  • HW: there’s an active market in insuring map makers against maritime mapping mistakes
    • it’s an expensive mistake when a supertanker runs aground…
  • GG: people have been hauled before the coastguards for using consumer mapping systems to navigate ships
  • GG: people tend to accept what they have on their phones

Should map data producers make end-user systems? Or leave it up to smaller companies?

  • IH: paper-based maps couldn’t change, so had to make one-off presentation decisions
    • now the data can change for the user context
  • CO: routing is hard - still dominated by big companies with background with navigation
    • but lots of space for smaller companies in other areas
  • IH: huge explosion of neo-geographers a few years ago
    • lots of mash ups and visualisations using geo data

What’s next in display? Will wearables move away from the concept of maps?

  • IH: we’ve always been shifting display methods…
  • Jo Rabin: years ago, a taxi driver in Chicago had a book of turn by turn directions but no map
  • HW: a large proportion of people get rather upset when they see a map (including Harry’s girlfriend…!)
  • Audience: worth thinking about how blind people navigate
  • IH: navigating a city by sound — turn away from the areas that sound threatening
  • HW: Google Street View seems a useful addition at the moment, but not a replacement for basic navigation
  • CO: 3D mapping is a bit gimmicky at the moment, but the data gathering behind it is massively useful
  • GG: for every mapping service, you only ever see less than 5% of the underlying data
  • CO: one of the best designed apps for mapping was an app for blind people on old Nokia phones — vibrate once for left, twice for right

What’s new and exciting in the next 3 years?

  • GG: indoor navigation
    • multiple companies working on it
    • but all currently incompatible
    • holy grail: directions from your house, through the tube to the exact exhibit in the V&A Museum
  • JT: putting open data together, e.g. being able to know when you’re in an area with a high risk of flooding after rain
    • 3D sim city-style simulation of London, with data such as crime rates superimposed
    • how can we bring in time-based information to our space-based navigation?
  • IH: early years education
  • IW: Ed Parsons (Maps @ Google) — “geospatial will no longer be geo-special”
    • not just a map screen in the app, but mixed throughout the app
    • promise of access to raw data
  • GG: NOAA have produced a low-power app which warns you about tornados, just using cell tower data
    • another innovative location-based service which doesn’t require a map…
  • CO: massive growth in transactional apps, e.g. Hailo & YPlan
    • also Google etc building an ecosystem e.g. their purchase of Waze
  • GG: wearables: I wouldn’t buy Google Glass, but I would buy a smart watch
  • GG: hospitals making savings by tracking their equipment locations (and stopping it sprouting legs)

Licensing

  • GG: businesses scared of using OSM data due to fear that they will have to share their private data
  • HW: share-a-like license aims to ensure that data is not taken and locked away
  • JT: OpenCorporates make company data available for free on a share-a-like license, but also offer a paid license with less viral terms
  • HW: OSM can’t do that — would have to agree with all contributors…

What about mobile operators? Indoor location?

  • GG: cell tower location works well in urban areas
    • less so in countryside…
    • and mobile operators in Germany are restricted to telling you only two of the nearby cell towers (not enough to get a fix)
    • mobile operators had their chance 8-10 years ago when they tried to sell their data to marketing companies
  • CO: starting to see traction in government
    • e.g. who’s moved on what roads at what speed
    • useful for real-time traffic management
    • high value areas: marketing & business intelligence

Coping with geospam…?

  • HW: Google are already dipping their toes in the water there (Ed: e.g. the map tiles have advertising built-in)
  • IH: walking past a shop and being offered a service
  • GG: been proved to be pretty worthless — you’ve already past when the message comes through
    • data of where you are not so valuable
    • where you’ve been is more valuable
    • where you will be is where the money will be made

Apple iBeacons — Bluetooth LE with an ID

  • CO: one inspiration for that comes from certain countries with restrictive regimes
    • walk into a bar and turn bluetooth on — get offered services that are normally not available
    • need a strong well-designed way to broker your data

Where are you failed by current map services?

  • HW: developing world — none of existing providers have coverage
    • Google Map Makers pushing hard
    • Open Street Map
  • Dan Appelquist: cycle maps for London
    • Google maps keeps on directing me through routes on Hampstead Heath that are not passable by bike
    • http://cyclestreets.net routing for bikes based on OpenStreetMap data
  • CO: pedestrian mapping
    • most of the data has come from driving cars around
    • huge value as people much more likely to make a purchase when they are walking
  • GG: but getting data is hard as many areas are private
    • no public right of way
    • hard for commercial companies to map
    • OSM has an advantage here — can’t really sue OpenStreetMap for trespass

How long rather than how far away?

  • IH: common GIS function, but doesn’t crop up in consumer stuff
  • JT: mapumental by mysociety — where should I live so that I can commute to work?

Royal Mail PAF file restrictions

big database of postcodes & business locations for the UK — really valuable and the Royal Mail are holding tight…

  • JT: campaigning to open it, but it seems extremely unlikely
    • would have to create an open street gazetteer
    • if someone would like to sponsor it — please see Jeni!

Linking to map model data

  • HW: most OSM data has a URL
  • GG: for roads — Nokia has Routing Segments, OSM has Ways, OS has ITN (Integrated Transport Network)

Summary

  • JT: collaboration is the most interesting area
  • IH: everybody should edit something on OpenStreetMap.org
  • GG: expected a question on location & privacy, Facebook & Google…
    • “I now trust big internet companies more than I do the government”

Next session

No comments: