Decentralised Mapping

A guide through the world of decentralized mapping projects

Bèr berkes Kessels

About this pres

  • Is online: berk.es/mapping-fission (git.webschuur.com/placebazaar/mapping-fission)
  • Uses the word Blockchain only once ☚
  • Uses the word Decentralised over Distributed, because Distribution means something different with mapping data.

About me

  • Bèr berkes Kessels
  • berkes (berkes.eth, mastodon, keybase, github and that birdsite)
  • Founder of placebazaar
  • Blog on berk.es
  • Mail on ber@berk.es

we’ll cover

  • Problems with mapping
  • Current solutions for mapping
  • What can decentralisation offer?
  • Deeper dive into one such solution: peermaps/ipfs.

Problems

Maps are expensive to make.

They require local knowledge. Feet on the ground.

Maps are always outdated.

The moment you publish a worldwide map, thousands of things have been moved, renamed, replaced, closed, opened and so on.

Maps are political and strategical.

Lots of pieces on a map are undebatable facts.

There is a park bench at X,Y.

Maps are also debatable.

This city is part of China/Israel/Thailand/Belgium

Maps are often strategical

  • Here’s a military base with N parking spots.
  • This is where all the speed camera’s are.
  • Here you can cross the border easily.

To summarise

  • are hard and expensive to make and maintain.
  • require local knowledge.
  • are politically motivated and censored.
  • are the last thing to put inside one monopolist.

Maps are the most important visualisation of data.

When did you last use a map.

Centralisation

Centralisation enables censorship

MCWV

Hampered by commercial interest

missing maps

… The ‘Wikipedia Of Maps’ Came To The Rescue

Enables privacy infringements

Google knows:

  • Where you work, live, sport, spend free time.
  • Where your boyfriend lives.
  • That you no longer visit your boyfriend.
  • When you quit your job.
  • Visit the liquor store more often.
  • Sleep on a bench.

Google Location services, and the “free” google maps app, are the biggest trojan horse. Deployed only to get the most valuable metadata about you.

Centralisation

  • Easy to censor.
  • Dependent on interest of central publisher.
  • Runs on users’ data (privacy)
  • Monopoly.
  • etc. The usual.

Current mapping

  • Google Maps
  • Open Street Maps
  • A bazillion GIS providers

example

There is mapping data.

  • Lots of it is open data.
  • Lots can be bought or licensed.
  • But it is spread all over.
  • And always in flux.

Solution

  • Decentralised collection of mapping data
  • Decentralised publication of mapping data
  • Entities and orgs glueing, and mixing all that data

Current Options

  • Google Business/maps
  • Apple Maps
  • Bing Maps
  • Open Street Map

Future options: Collection

Future options: Distribution

IPFS by placebazaar
  • Streaming

GTFS example

Dynamics of mapping data

  1. Geography: continents, mountains, lakes
  2. Physical structure: Roads, bridges, location of cities
  3. Political: Boundaries, borders
  4. Metadata: Speed limits, turn restrictions, park benches, restaurant
  5. Logistics: traffic, transport

Obstacles

  • Collection of data is complex.
  • Data is continuously updating.
  • Data is HUGE.
stats

Data is never “finished”. Or 100% dependable.

Errors
Vandalism
Vandalism

Data is everywhere (but unfindable).

example { target=_blank }

Standards

XKCD 927 Standards

Standards: just one example

  • Netherlands requires Rijksdriehoekstelsel
  • How many official Spatial reference systems are there? Guess?
  • Nope: 4362 (in EPSG)

Mixing, Mashing and Merging is a problem.

Decentralised

Just because it is possible to build distributed systems does not necessarily mean that it is a good idea. — Andrew S. Tanenbaum in Distributed Systems: Principles and Paradigms

Solutions by centralising

  • Weekly/Nightly imports: just rebuild the database.
  • Event stream, handle changesets. example live updates

Solutions by centralising

  • Central Quality Assurance (tools)
  • Validation by crowd
  • Effort orchestration
  • Standards enforcement and -emergence.

Solutions through decentralisation

Breaking Monopolies?

  • There’s enough data. Open and Free.
  • Data gathering of map-data is decentralised by nature.
  • (Well there’s satellite and AI).
  • We need more interfaces around existing datasets. The read-side is not distributed.

Breaking Censorship?

  • On the read-side.
  • Developers need more options (than Google maps).
  • Developers need easier access to datasets.
  • PlaceBazaar is building a PoC OpenOpeningsTijden; a business directory

Taking back ownership?

  • Any personal location data must be owned by users.
  • Any other data must be open data.

What can the decentralisation-movement do?

Make apps.

(and make them privacy-first)

Use apps.

OSMand, Maps.me, Or one of the 160+ android or 126 iOS apps using open data maps.

Use free mapping tools

  • Display maps: Leaflet, OpenLayers
  • In apps: Mapbox SDK, Skobbler, Nutitek, and many more
  • Run your own tileserver: Mapnik and many more.
  • Tune your own routing: GraphHopper, OpenTripPlanner, Brouter, Openrouteservice
  • And more usages

Release open data (yay! even more data)

Peermaps

example: luxembourg

Storage

Example of the file structure

R-tree

R-tree

Code.

ssh kind-jefferys.webschuur.com
peermaps data 6.3281,49.4440,6.6800,49.6160 | bat

diff -u <(peermaps files 6.3281,49.4440,6.6800,49.6160) <(peermaps files 6.3281,49.0,6.6800,49.6160)


peermaps data 6.3281,49.4440,6.6800,49.6160 | grep "fuel"

peermaps data 5.4080,49.4137,7.5641,50.2235 | grep "fuel"