This document introduces ESTA-LD, a tool for exploring and analyzing spatio-temporal linked statistical data on the web. It summarizes ESTA-LD's key capabilities, which include visualizing statistical data on charts and maps, filtering and selecting dimensions like time and location, and aggregating data. The document also discusses how ESTA-LD models spatio-temporal data using the RDF Data Cube vocabulary and SDMX standards. Future work aims to expand ESTA-LD's chart types, analysis functions, and reduce data replication across datasets.

1. Making the web an
exploratory place
for geospatial data
1
ESTA-LD: Exploring Spatio-Temporal Linked
Statistical Data
Vuk Mijović,Valentina Janev, Dejan Paunović
Mihajlo Pupin Institute

2. Outline
• Introduction
• Modelling
• ESTA-LD
• FutureWork
1 5 S e p t e m b e r 2 0 1 5 2

3. Motivation
• OGD (Open Government Data) initiatives have helped to open public data on:
– transport
– education
– infrastructure
– health
– environment
– …
• Public sector information is usually statistical in nature:
– SORS (Statistical Office of the Republic of Serbia
– SBRA (Serbian Business Registers Agency)
– Eurostat
• Public sector information often spans
across space and time
– Regional development
– Tourism data
– GDP
– …
• Statistical data on the web
– Create mash-ups
– Create visualizations
1 5 S e p t e m b e r 2 0 1 5 3

4. Scenario
1 5 S e p t e m b e r 2 0 1 5 4
Agents –Data providers
Services
LOD
convertor
LOD
convertor
Harvesting
Sharing metadata
with Joinup
Metadata catalogue
SPARQL
endpoint
SPARQL
endpoint
RDBMS RDBMS
• Spatio-temporal analysis
• Querying and Exploration
• Authoring of Linked Data
• Semi-automatic link discovery
• Enrichment and Repair
• Extraction and Loading

5. RDF Data CubeVocabulary
• Publishing multi-dimensional (statistical) data in a way that it can
be linked to related data sets and concepts
• Compatible with SDMX
• W3C Recommendation
• Kinds of data:
– Observations
– Organizational struct.
– Structural metadata
– Reference metadata
• The cube model:
– Dimensions
– Attributes
– Measures
1 5 S e p t e m b e r 2 0 1 5 5

6. Data Cube Example
1 5 S e p t e m b e r 2 0 1 5 6

7. Modelling Spatio-Temporal Data
• Use SDMX COG (ContentOriented Guidelines) available in RDF:
– Temporal dimension: sdmx-dimension:refPeriod
– Spatial dimension: sdmx-dimension:refArea
• Temporal Dimension
– Concept: sdmx-concept:refPeriod
– Time role: sdmx:TimeRole
– Range: xsd:gYear, xsd:gYearMonth, xsd:date…
• Spatial Dimension
– Concept: sdmx-concept:refArea
– There is no spatial role
– Define geometries for concepts using geo:hasDefaultGeometry and
geo:asWKT
– Hierarhies:
• Define a code list
• Define hierarchy levels using the SKOS vocabulary (skos:brader and skos:narrower)
1 5 S e p t e m b e r 2 0 1 5 7

8. Example: Spatial andTemporal
1 5 S e p t e m b e r 2 0 1 5 8

9. ESTA-LD
• Enables exploration and
analysis of spatio-temporal
linked statistical data
• Based on the RDF Data
Cube vocabulary
• Works with any SPARQL
Endpoint
• Part of the Linked Data
Stack developed within EU
FP7 project GeoKnow
1 5 S e p t e m b e r 2 0 1 5 9

10. Setting Parameters
• Choose graph and dataset
• Choose measure, set dimension values and select dimensions to visualize
1 5 S e p t e m b e r 2 0 1 5 1 0

11. ChartVisualization
• Visualization of up to
two dimensions
• SwitchingAxes
• Stacking
• Swapping series and
categories
1 5 S e p t e m b e r 2 0 1 5 1 1

12. Temporal Dimension
• Time Chart:
– Dedicated chart for the
time dimension
• Select a period
• Slide through time
– Works with
SBRA/SORS code list
and XSD types
– Selection of the time
window updates the
map
1 5 S e p t e m b e r 2 0 1 5 h t t p : / / g e o k n o w. e u 12

13. Spatial Dimension
• Choropleth Map:
– Visualizes data on a map
– Supports hierarchies
• Regions, municipalities …
– Selecting a region updates chart
– Regions need to point to polygons
supplied asWKT strings
– In combination with the time chart,
regions are coloured:
• Based on the selected time window,
• Based on every possible time
window of the same duration as the
one selected.
1 5 S e p t e m b e r 2 0 1 5 h t t p : / / g e o k n o w. e u 13

14. AggregatedColoring
1 5 S e p t e m b e r 2 0 1 5 1 4

15. Support for Preparing Data Cubes
• Temporal Dimension:
– In most cases where time is represented with URIs, those URIs contain year/month/date.
– Provides transformation of URIs to xsd types based on the provided pattern
• Spatial Dimension:
– In most cases country name or code is embedded in the URI
– Search for polygons in the LinkedGeoData dataset and add them to the Cube
1 5 S e p t e m b e r 2 0 1 5 h t t p : / / g e o k n o w. e u 15

16. Final Remarks
• Try ESTA-LD:
– GitHub: https://github.com/GeoKnow/ESTA-LD
– Demo: http://geoknow.imp.bg.ac.rs/ESTA-LD
• Future Work:
– Add more chart types
– Enable comparisons, merging, slicing, etc.
– Reducing the need for replication (DSDs, polygons…)
– Advanced search and filtering capabilities
1 5 S e p t e m b e r 2 0 1 5 1 6

17. Making the web an
exploratory place
for geospatial data
17
Thank you for your attention!
Questions?