DataShine Automated Thematic Mapping of 2011 Census Quick Statistics Oliver O Brien, James Cheshire Department of Geography University College London UKDS Census Applications Conference, July 2015
Contents Summary of the DataShine project Cartographic considerations Technology considerations Advanced User Features Coming soon to DataShine Future directions for web mapping of census data
DataShine is A series of websites Map social science geodemographic datasets Initially using data from (or derived from) the 2011 UK Census The focal point of each website is a full-window map Designed to be data maps that looks like regular maps. Quick Statistics 60 Tables 1500+ Categories Geodemographic Classifications 2 Tables 15 Categories Travel-to-Work Flows 1 Table 11 Categories Election Results 2 General Elections 8+ Categories
DataShine is A discovery tool Uses a simple user interface which invites exploration of the map Uses vivid colours to emphasise variation Numeric data display kept discrete but displayed on mouseover
DataShine is A display tool Uses unique URLs so particular views can be easily shared Boundaries (e.g. local authorities) can be overlaid (KML, GeoJSON) Current view can be printed with PDF creator (or via browser) Suprisingly hard to implement printing/display of slippy maps (in 2014) Now much easier with the release of OpenLayers 3.X and Canvas printing (2015)
DataShine is used by Interested members of the public Local authority researchers Campaign groups The press
DataShine: The Website (datashine.org.uk)
DataShine is Based entirely on open data Census data from the Office for National Statistics (+ NISRA + GROS) Contextual data from Ordnance Survey Open Data (& OpenStreetMap) ONS & OS data (now) licenced under the Open Government Licence
Cartographic Techniques in DataShine: 1 Technique 1: We create a custom semi-transparent maps, using street networks, town/city names and major natural features We then underlay choropleths on this map We use ColorBrewer colour ramps only Context layer doesn t use colours (except slight accents for water and land) Single font used, buffered text, character spacing increased where appropriate
Cartographic Techniques in DataShine: 2 Technique 2: We constrain the data to weighted centroids or lines, rather than areas We can then use a regular background map, e.g. from OpenStreetMap.org, Microsoft, HERE Maps, ArcGIS REST API N.B. Not Google Maps any more can now only be used with their own API.
DataShine: Detailed Cartography
DataShine: Detailed Cartography
DataShine: Technical Considerations Vector vs Raster Display Vector allows for maximum flexibility Bandwidth considerations Simplification strategies: client vs server Hybrid: Raster tiles, vector metadata Caching Spatial indexing
DataShine: Technical Considerations Context: Pre-rendered 1 layer x 8 zoom levels x up to 2 million tiles Choropleth: On the fly 1600 layers x 8 zoom levels x up to million tiles x 15 colours x custom colour ramp range Vector metadata: On the fly (but could be prerendered) 1600 layers x 8 zoom levels x up to 2 million tiles x custom colour ramp range
DataShine: Technical Considerations Vector metadata UTF Grid format Mapping API calculates current grid cell that the mouse is over and passes the corresponding metadata to the web app
DataShine: Technical Considerations
DataShine: Technical Considerations Simplification and Optimisation Considerations We use two geographies Ward and OA Can cause MAUP Reduces speckle Faster rendering times for simpler geographies Flat file for the table range avg/s.d. stats Avoids initial database query upon simple loading of the website Every click loses 50% of the audience (maybe?) Tiles requests hit cache then database if not available Local area rescaling and data downloads go straight from server PostGIS database with GiST spatial indices
DataShine: Technical Considerations Data catagorising strategy Based entirely on mean and S.D. Always uses 8 colours Want to minimise chance of < 0% and > 100% bins
Coming Soon July 2015: DataShine Scotland Census & Commute (at LA level) July 2015: Open APIs Choropleth XYZ Tiles (OTF) Context Layer XYZ Tiles (pre-made) Download the whole-area aggregate data Bounding-box data download Open-sourcing the DataShine JavaScript code
Coming Soon September 2015: IMD/IDACI 2015 mapping Also Land Registry mapping Exploring possibility of incorporating other countries census data in the same interface e.g. US 2010 Census SF1 data May create/colour a road-network instead of building outlines, or use OSM data (but this is not complete)
DataShine Scotland
Future directions for web mapping of census data Vector-based e.g. Mapnik Vector Tile (MVT) format Tricky to minimise file size/processing time without compromising on map quality. Mobile-focused Redesigning for small screen Current-area focused Current view easily shareable on social media withimagery
Future directions for web mapping of census data Seamless linking of census and non-census data Single link to current view Social media permeability is crucial
Website Addresses & Contact The DataShine Team is: James Cheshire (@spatialanalysis) Oliver O Brien (@oobr) DataShine http://datashine.org.uk/ The DataShine Blog http://blog.datashine.org.uk/ Email: o.obrien@ucl.ac.uk