A bit about me....

• Studied a Masters at UCL
• Went on to be a developer evangelist at Esri UK
• Started working for the Ordnance Survey just over a year ago
• I am leaving next month to work for a startup!

• UKs National Mapping Agency
• Formed in 1791
• Granular, current mapping data for the UK
• 500+ million geographical features in our database

Geovation

Purpose

• Bit of history about the geo industry
• Overview of geo technology
• Libraries and frameworks for handling geo data

I'll start with a story time

Jon Snow was one of the first people to analyse geo data to solve a problem

Over time this field was formalised into an subject entitled 'Geographic Information Science/Systems' (GIS)

GIS are systems for organising, storing, retrieving, editing and analyzing spatial data

Wrangling spatial data can have many outputs, but the most common is maps

But, we can also do other things with data other than plot it

GIS is often considered to have formalised in the 1960s/1970s with the advent of modern computing

New Tools

Mobile and sub 100m positioning 

Source: Andreessen Horowitz, 2015

We can approximate that about 70% of smartphone devices have a GPS chip

Lots of startups have embraced spatial data

Embracing Services and APIs

The Geospatial Stack

Firstly, Data Formats

• Shapefile
• GeoJSON
• KML
• GML
• CSV (points!)

All parts of the architecture will need to be spatially capable 

Spatial Architectures

Lets start at the backend

Understanding Spatial Web Services

• Maps
• Geocoding
• Reverse Geocoding
• Routing (A to B, Travelling Salesman etc)
• Points of Interest
• Boundaries

Services

Roll Your Own

• Bring your own data
• Maintain your own infrastructure
• More granular control
• Requires knowhow
• Pay for infrastructure and dev time

• Generally has content
• Generally just use simple web interfaces (i.e. REST)
• Docs may vary
• Pay per request or annual/monthly fee

Spatial Web Services

Spatial Databases

SELECT ST_Intersection(s.geom, r.geom), s.state, r.river
FROM usa_states as s, usa_rivers as r
WHERE s.state = 'Texas'

Spatial databases leverage spatial indexes

How might a spatial service look under the hood?

What about the frontend?

Open Source Web Mapping Libraries

Fast, small, consistent API, good docs

Comprehensive, powerful, longstanding

Niche, more oriented around 'data viz' community

// Create map in our page, with a div with an id of mapid
var mymap = L.map('mapid').setView([51.505, -0.09], 13);

// Add a basemap
L.tileLayer('https://api.tiles.mapbox.com/v4/{id}/{z}/{x}/{y}.png', {
    attribution: "Insert mapbox attribution here"
    maxZoom: 18,
    id: 'your.mapbox.project.id',
    accessToken: 'your.mapbox.public.access.token'
}).addTo(mymap);

// Example of adding a marker
var marker = L.marker([51.5, -0.09]).addTo(mymap);

// Example of some events
function onMapClick(e) {
    alert("You clicked the map at " + e.latlng);
}

mymap.on('click', onMapClick);

A more complex Leaflet example

Closed Source Web Mapping Libraries

Well known, lots of questions online, hard to understand T&Cs

Comprehensive, powerful, hard to understand T&Cs

What if I just want to do stuff on my machine?

Local Tools

Python

• Shapely
• Fiona
• GeoPy
• Ogr/GDLA Bindings
• PyQGIS
• ArcPy (closed)

R

Questions?