Data Visualisation

Prof. Keith Andrews

IICM, Graz University of Technology, Austria

http://www.iicm.tugraz.at/keith/

Slides:
http://keithandrews.com/talks/2012/epsi2012/

ePSI Platform Conference, Fri 16 Mar 2012, Rotterdam, Netherlands.

Open Data Formats

• Spreadsheet
  • Excel (xls), OpenOffice (ods), comma-separated value (csv), JSON, etc.
  • machine-readable
  • (potentially) with a guaranteed URI
• SPARQL + RDF
  • RDF triples in say XML notation.
  • machine-readable
  • (potentially) with a guaranteed URI
  • (potentially) live data = always current

Tools for Massaging Data

• Scripting Languages
  • Python
  • Perl
• Data Cleaning
  • Wrangler http://vis.stanford.edu/wrangler/
    (web service)
  • Google Refine http://code.google.com/p/google-refine/
    (download, but runs locally in web browser)

Tools for Visualising Data

Four varieties:

• Web Services
  • must be online
  • point and click
  • can (sometimes must) share data
  • can (sometimes must) share finished visualisations
• Standalone Applications
  • download, install, and use locally
  • point and click
  • generally produce static visualisations (limited interaction)
• Scripting Languages (with Libraries)
  • download, install, and use locally
  • programming
  • generally produce static visualisations (limited interaction)
• Toolkits (with Programming APIs)
  • download, install, and use locally
  • even more programming
  • can produce fully interactive visualisations

Web Services

• IBM Many Eyes
  • upload data (always public)
  • create public visualisations (Java applets)
  • http://many-eyes.com/
• Google Spreadsheets + Gadgets
  • upload data (private or shared)
  • insert gadgets (private or shared)
  • custom gadgets (Google Vis API) [Liquid Diagrams Gadgets]
  • http://docs.google.com/
• Tableau Public
  • upload data (public)
  • analyse data
  • create visualisations (public)
  • create dashboards (public)
  • http://tableausoftware.com/public/
• Tibco Silver Spotfire http://silverspotfire.tibco.com/
• verifiable.com (dead)
• swivel.com (dead)
• graphmonster.com (coming soon) http://graphmonster.com/
• datamaps.eu http://datamaps.eu/
• Google Maps API

Standalone Applications

• Excel
• OpenOffice Calc
• Adobe Illustrator
• Liquid Diagrams (not yet released)

Scripting Languages (with Libraries)

• R
  • lattice
  • ggplot2
  • portfolio
• Python
  • SciPy http://www.scipy.org/
  • Matplotlib http://matplotlib.sourceforge.net/
  • Veusz http://home.gna.org/veusz/
  • SciDAVis http://scidavis.sourceforge.net/
  • orange http://orange.biolab.si/
• Processing http://processing.org/

Toolkits (with Programming APIs)

• Java
  • Prefuse http://prefuse.org/
  • InfoVis Toolkit (ivtk) http://ivtk.sourceforge.net/
• Actionscript (Flash, Flex)
  • Flare http://flare.prefuse.org/
  • Kap Lab Kalileo http://lab.kapit.fr/default/Kalileo/
  • birdeye http://code.google.com/p/birdeye/
  • Modest Maps http://modestmaps.com/
• Javascript (SVG, HTML5, CSS3)
  • ProtoVis http://protovis.org/
  • Javascript InfoVis Toolkit (jit) http://thejit.org/
  • D3 http://mbostock.github.com/d3/
  • Polymaps http://polymaps.org/
• Many, many others (for C, C++, ...)

Tools for Visualising Data Given in RDF

• Not many...
• UK Archives; Cabinet Office Organogram; http://data.gov.uk/organogram/cabinet-office
http://code.google.com/p/linked-data-api/source/browse/branches/organogram-2011-03-31/
  (Server-side processing of RDF, ships JSON to browser, and uses jit for visualisation).

References

• Nathan Yau; Visualize This; Wiley, Jul 2011. 0470944889
• Noah Iliinsky and Julie Steele; Designing Data Visualizations; O'Reilly, Sep 2011. 1449312284
• Matthew Ward, Georges Grinstein, and Daniel Keim; Interactive Data Visualization; A. K. Peters, Jul 2010. 1568814739

Conferences

• EuroVis 2012; 05-08 Jun 2012, Vienna, Austria. http://eurovis.org/
• OGD Austria 2012; 26 Jun 2012, Linz, Austria. http://ogd2012.at/
• VisWeek 2012; 14-19 Oct 2012, Seattle, WA, USA. http://visweek.org/

Thank You

Slides:
http://keithandrews.com/talks/2012/epsi2012/

Extra Slides

A Eurostat Vis Tool

1. Get RDF/XML data from Eurostat wrapper
2. Query data using SPARQL
3. Generate JSON format
4. Visualise using ProtoVis

Example Eurostat Data in RDF/XML

<qb:Observation>
<qb:dataset rdf:resource="/id/tin00090#ds"/>
<geo rdf:resource="/dic/geo#AT"/>
<dcterms:date>2003</dcterms:date>
<sdmx-measure:obsValue>48</sdmx-measure:obsValue>
</qb:Observation>
...

SPARQL RDF query

<html>
<script src="https://ajax.googleapis.com/ajax/libs/jquery/1.6.1/jquery.min.js"
type="text/javascript"></script>
<script src="./dist/js/jquery.rdfquery.core-1.0.js"
type="text/javascript"></script>
<script src="./dist/js/jquery.rdfquery.rdfa-1.0.js"
type="text/javascript"></script>

<head>
<title>RDF 2 JSON<title>
</head>

<body>
  <p>Querying Eurostat data using SPARQL</p>
  <script>
  $.get('./RDFexamples/tin00089_test2.rdf', null, function (rdfXml) {
    var rdf, json = {};
    var countries = new Array();
    var values = 
    rdf = $.rdf()
      .load(rdfXml)
      .prefix('', 'http://ontologycentral.com/2009/01/eurostat/ns#')
      .prefix('qb', 'http://purl.org/linked-data/cube#')
      .prefix('rdf', 'http://www.w3.org/1999/02/22-rdf-syntax-ns#')
      .prefix('dcterms', 'http://purl.org/dc/terms/')
      .prefix('sdmx-measure', 'http://purl.org/linked-data/sdmx/2009/measure#');
      rdf
      .where('?observation a qb:Observation')
      .each(function () {

...

JSON-like Data Example

var eu_stats = {
'AT': {name:'Austria', a:[33, 46, 54, 58, 64]},
'BE': {name:'Belgium', a:[48, 56, 60, 63, 70]},
'BG': {name:'Bulgaria', a:[10, 15, 21, 26, 26]},
'CY': {name:'Cyprus', a:[12, 20, 33, 47, 51]},
'CZ': {name:'Czech Republic', a:[17, 28, 36, 49, 54]},
'DE': {name:'Germany', a:[34, 50, 55, 65, 75]},
'DK': {name:'Denmark', a:[63, 70, 74, 76, 80]},
...
'SI': {name:'Slovenia', a:[34, 44, 50, 56, 62]},
'SK': {name:'Slovakia', a:[11, 27, 35, 42, 49]},
//'TR': {name:'', a:[0, 17, 22, 26, 34]},
'UK': {name:'United Kingdom', a:[44, 57, 62, 69, 0]}
};
eu_stats.maxYear = 2010;
eu_stats.minYear = 2006;
eu_stats.yearIdx = function(year) { return -2006+year; }

Final Visualisation