From ImageWeb

• Back to Spider_Project
• See also PLOS_Paper

Contents 1 Key terms - automatic extraction 2 Heatmap Overlay - Gmaps mashup 3 Heatmap Overlay - Gmaps + data from another article 3.1 Issues 4 Fig3 overlay tool 5 Leptospira Age data mashup 5.1 Data from 5.2 Issues 5.2.1 Comparable X axis 5.2.2 Comparable Y axis 5.2.3 Power of studies 6 Citation Networks 6.1 Visualisation 6.1.1 Graphgear 6.1.2 Welkin 6.2 Data sources 7 Lively Data 7.1 Visualisation 7.2 Extraction from images 8 Field studies Geo mashup 8.1 Visualisation 8.2 Time-based visualisation

Key terms - automatic extraction

• Opencalais http://imageweb.zoo.ox.ac.uk/pub/2008/miningdev/calais/calais.htm
• uBio http://imageweb.zoo.ox.ac.uk/pub/2008/miningdev/ubio/index.htm

Heatmap Overlay - Gmaps mashup

http://maps.google.com/maps?q=http://imageweb.zoo.ox.ac.uk/pub/2008/geomashupdev/overlay/heatmap_leptosipra_paudalima_salvador_brazil.kml

• Fig 3 in original context
• Story? This visualisation puts the location of the study into geographical context- especially if you zoom out
• As a side effect, we have these machine readable statements

Heatmap Overlay - Gmaps + data from another article

http://maps.google.com/maps?q=http://imageweb.zoo.ox.ac.uk/pub/2008/geomashupdev/overlay/leptospira_heatmap_mashup.kml

• This is a mash up of our fig 3 with with fig 1 from this other Ko paper.
• This mashup shows that the slum study site in our plos paper is in an area of overall high incidence of leptospirosis, according to the census.

Issues

• Overlay imperfect - coastal outline of fig 1 does not perfectly correspond with gmaps.

Fig3 overlay tool

http://imageweb.zoo.ox.ac.uk/pub/2008/datavisualisationwidgets/overlay/fig3/

• The maps are draggable and opaque so you can align them as you wish.
• It is interesting to see how the large trash piles (red blobs) align with the high incidence of lept. If you line all 3 up one thing which becomes evident is that data which would be really interesting is missing from this study . The top white "arm" is a high hill, it also has quite a few trash piles on it, but it was not surveyed (black line = study area). If I could annotate this enhanced paper, I would ask Ko why they did not survey the houses on the hill. (NB clearly there *are* houses on the hill - I can see them here
• User story? As a researcher, I want to better evaluate the (extent of?/quality of?) data presented in this paper.

Leptospira Age data mashup

http://imageweb.zoo.ox.ac.uk/pub/2008/agemashup/leptospira-age-mashup_v2.xls

Data from

1. Ko et al (our Plos paper) http://www.plosntds.org/article/info:doi%2F10.1371%2Fjournal.pntd.0000228 2. Libraty et al (paper B) http://www.plosntds.org/article/info:doi%2F10.1371%2Fjournal.pntd.0000111 3. Maciel et al (paper D) http://www.plosntds.org/article/info:doi%2F10.1371%2Fjournal.pntd.0000154

Issues

Comparable X axis

• Ranges non-aligned.
• Data non/barely overlapping.

Comparable Y axis

• Since Libraty did not include control data for each age category, the only measure available was proportion of leptospira patients in this age category (ie, over the total). The other two data sets were transformed into comparable values.

Power of studies

• 489 patients in Ko study. 12 and 22 in the other studies. Should get equal prominence?

Citation Networks

Visualisation

Graphgear

http://imageweb.zoo.ox.ac.uk/pub/2008/citationnetworkdev/visualisation/graphgear/

Welkin

http://imageweb.zoo.ox.ac.uk/pub/2008/citationnetworkdev/visualisation/rdfcitationdata/Ganoza_2006_DeterminingRisk_extra1.rdf / Welkin. See Citation Networks for more info.

Data sources

See Citation Networks. (Crossref, GScholar)

Lively Data

Visualisation

• OAT Charting Tools : http://imageweb.zoo.ox.ac.uk/pub/2008/datavisualisationwidgets/oat/fig2/

Extraction from images

• ocropus (webservice) Tested on Table 1. Results in plos\Trunk\mining\ocropus\OCR Output.txt. Extracted text successfully, but lost the formatting.

Field studies Geo mashup

Visualisation

• Demo: http://maps.google.com/maps?q=http://imageweb.zoo.ox.ac.uk/pub/2008/geomashupdev/leptospirademo/leptospira_field_studies.kml
• Our code consists of a KML file. This is passed as a parameter to google maps which handles all the interactive features in this demo.
• SVN:

  PLOS/Trunk/geomashup/leptospirademo/

• See also Geo Mashup eg Yahoo Maps alternative visualisation, geocoding options for creating the data, etc.

Time-based visualisation

• Demo: http://imageweb.zoo.ox.ac.uk/pub/2008/geomashupdev/leptospirademo_time/
• Inspired by http://www.mapstraction.com/demo-filters.php
• Our code consists of:
  • KML file - as above with addition of temporal information KML API
  • Index.html and gmapsjscript.js. Google maps can not currently handle any temporal features and therefore this demo is a hosted page with javascript. Functionality:
    • 1. KML parsing routines - we have to explicitly read the kml to get at the timestamp/when element. Bespoke code (no libraries avail).
    • 2. Time-based "filters" on markers - uses mapstraction library.
    • 3. Slider - uses scriptaculous library. (TODO swap to yui?)
• SVN:

  PLOS/Trunk/geomashup/leptospirademo_time/