2014-03-06-collaborative-geospatial-data.html

---
layout: reveal
title: Collaborative Geospatial Data
description: Open glue for the web of data 
author: Stuart Harrison
twitter: pezholio
---

{% include odi_logo.html %}

{% include standard_title.html %}


<style type="text/css">
@import url(https://fonts.googleapis.com/css?family=Quicksand:400,700,400italic,700italic);
@import url(https://fonts.googleapis.com/css?family=Open+Sans:400italic,700italic,400,700);

reveal {
  font-family: "Open Sans", sans-serif;
  font-size: 36px;
  font-weight: 200;
  letter-spacing: 0.6em;
  color: #333333; }

.reveal h1,
.reveal h2,
.reveal h3,
.reveal h4,
.reveal h5,
.reveal h6 {
  margin: 0 0 20px 0;
  color: #333333;
  font-family: "Quicksand", sans-serif;
  line-height: 1.2em;
  letter-spacing: -0.08em;
  text-transform: uppercase;
  text-shadow: none;
 }
 
 .reveal img {
  border: none !important;
 }
 
 .organisations img {
  margin: 0 10px !important;
 }
 
</style>


<section data-background="2014-03-06-collaborative-geospatial-data/mappa-mundi.jpg">
	<h2>Geo data is brilliant</h2>
	<ul>
	  <li>Easily reference-able</li>
	  <li>Great for visualisations</li>
	  <li>Even better for linking datasets together</li>
	</ul>

  <aside class="notes">
    <p>As we all know, Geo data is brilliant.</p>
  </aside>
</section>

<section>
	<h2>HOWEVER</h2>
	
	<aside class="notes">
	 <p>However, in the UK, we have a problem</p>
	</aside>
</section>

<section>
	<h2>Many organisations</h2>
	
	<div class="organisations">
	<img src="2014-03-06-collaborative-geospatial-data/os.png" />
	<img src="2014-03-06-collaborative-geospatial-data/royal-mail.png" />
	<img src="2014-03-06-collaborative-geospatial-data/land-registry.png" />
	<img src="2014-03-06-collaborative-geospatial-data/lichfield-district-council.jpg" />
	<img src="2014-03-06-collaborative-geospatial-data/national-grid.png" />
	</div>
	

  <aside class="notes">
    <p>Geodata is collected by a wide range of organisations, who don't always use the same standards.</p>
    <p>And, more crucially..</p>
    
  </aside>
</section>

<section data-background="2014-03-06-collaborative-geospatial-data/padlock.jpg"> 
	<h2>Closed datasets</h2>
	
	<div class="photo-credit">
      <a href="http://www.flickr.com/photos/simon_cocks/4534585117">Padlock</a> by
      <a href="http://www.flickr.com/photos/simon_cocks/">Simon Cocks</a>
      <a href="http://creativecommons.org/licenses/by-nc-sa/2.0/">(CC BY-NC-SA 2.0)</a>
  </div>

  <aside class="notes">
    <p>Despite some of the great strides that OS have made in this area, much of this data is closed</p>
  </aside>
</section>

<section data-background="2014-03-06-collaborative-geospatial-data/Tim_Berners-Lee_in_hi_viz_vest.jpg">
	<h2>Enter OpenStreetMap</h2>
	
	<ul>
	  <li>Not appropriate for temporary information (events etc)</li>
	  <li>No direct access to raw data via URIs</li>
	  <li>Sharealike licence</li>
	  <li>Centralised</li>
	</ul>
	
	<div class="photo-credit">
      <a href="http://wiki.openstreetmap.org/wiki/File:Tim_Berners-Lee_in_hi_viz_vest.jpg">Tim Berners-Lee in hi viz vest</a> by
      <a href="http://wiki.openstreetmap.org/wiki/User:Ivansanchez">Ivansanchez</a>
      <a href="http://wiki.openstreetmap.org/wiki/OpenStreetMap_License">(CC BY-SA 2.0)</a>
  </div>

  <aside class="notes">
    <p>Thankfully, there is help in the form of OpenStreetMap.</p>
    <p>BUT</p>
    <p>Temporary information can't be added to geodata (for example, events in a park wouldn't be appropriate)</p>
    <p>It's not part of the linked data web, so doesn't provide direct access to the raw data via URIs</p>
    <p>The sharealike licence can put people off using it</p>
    <p>Storing geospatial data in one centralised service may not be the best fit with a distributed web architecture</p>
  </aside>
</section>

<section>
	<h2>The solution(s)?</h2>
	
	<ul>
	  <li>Capture existing openly licensed data</li>
	  <li>Extract into a common format</li>
	  <li>Crowdsource the gaps</li>
	</ul>

  <aside class="notes">
    <p>How do we get around these issues?</p>
    <p>Thankfully there's lots of open data available on the web with geodata included</p>
    <p>So, we can hoover all this up, and provide a platform to crowdsource it</p>
  </aside>
</section>

<section>
	<h2>Open Green Spaces</h2>
	
	<ul>
	  <li>Taking green space data from various sources</li>
	  <li>Allowing the public to add extra context and locations</li>
	  <li>Licensed under the UK Open Government Licence</li>
	</ul>

  <aside class="notes">
    <p>Green space data is very disperate, and held by local councils, NGOs and community groups</p>
    <p>This platform would allow people and organisations to add information about local parks and open spaces</p>
    <p>Either individually, or by releasing their own data under the OGL</p>
    <p>Data would be released in a common format under the OGL</p>
  </aside>
</section>

<section>
	<h2>Open Addresses</h2>
	
	<ul>
	  <li>Building upon Open Green Spaces</li>
	  <li>Taking existing OGL-licensed content:
	    <ul>
	      <li>Land Registry price paid data</li>
	      <li>Companies House</li>
	      <li>OS Locator</li>
	      <li>CodePoint Open</li>
	    </ul>
	  </li>
	</ul>

  <aside class="notes">
    <p>The second step will build on the technology used by Open Green Spaces to provide a crowdsourced address platform</p>
    <p>Taking existing OGL-licensed content, such as...</p>
    <p>And get the data into a common format</p>
  </aside>
</section>

<section>
	<h2>Closing the gaps</h2>
	
	<ul>
	  <li>Programatically guessing gaps (e.g. house numbering)</li>
	  <li>Crowdsourcing sources of data from other organisations</li>
	  <li>Allowing individual edits (similar to OSM)</li>
	  <li>Confidence scoring: How reliable is this data point?</li>
	</ul>

  <aside class="notes">
    <p>We could probably get 50% of the data from existing sources, but how do we get the rest?</p>
    <p>We can guess gaps in (for example) house numbering</p>
    <p>We can allow bulk uploads from organisations such as local authorities, housing associations (subject to licensing)</p>
    <p>We can allow individual edits</p>
    <p>All of this will then get a confidence score (e.g. guessed data would have a lower confidence than data from the Land Registry)</p>
  </aside>
</section>

<section>
	<h2>The future</h2>
	
	<ul>
	  <li>Stage 0: <a href="https://github.com/theodi/parse-uk-addresses">https://github.com/theodi/parse-uk-addresses</a></li>
	  <li>Stage 1: Green Spaces</li>
	  <li>Stage 2: Funding?</li>
	</ul>

  <aside class="notes">
    <p>We've already done some work, and we're planning to start work on the green spaces project ourselves</p>
    <p>BUT, we need funding to take it to the next level</p>
    <p>Get in touch!</p>
  </aside>
</section>

{% include odi_tech_team.html %}