Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Spatial-temporal data visualization in STempo

Spatial-temporal data visualization in STempo

Alice Rühl, Penn State University
Scott Pezanowski, Penn State University
Frank Hardisty, Penn State University

The increasing importance of Visual Analytics, particularly in response to the advent of 'big data', calls for greater attention to maps in analytic tools and applications. While maps provide a familiar, understandable interface to underlying data, not enough has been done in the field of Visual Analytics to advance the design and usage of maps.

This study explores the development of map components within multi-view applications using multivariate and zoom-level dependent representations of news event data. Integrated in the system with a timeline, a word cloud and other view components, we consider the traditional 'rules' of cartography as well as aspects of human-computer interaction and interactive design research. This allows users to efficiently explore complex spatio-temporal data. Our approach was validated by re-developing the mapping component inside STempo, a project developed at the GeoVISTA Center, Pennsylvania State University.

Nathaniel V. KELSO

October 09, 2014
Tweet

More Decks by Nathaniel V. KELSO

Other Decks in Design

Transcript

  1. Outline • STempo project - http://www.geovista.psu.edu/stempo/ o Use and User

    o Event Data o Technology • Base map • Data Visualization and Interactions • Future Work • Demonstration
  2. STempo - Use • Identification and visualization of potentially interesting

    spatial event sequences (Robinson et al. 2014) • Explore known patterns, but also understand previously unknown patterns • Discover significant patterns, drill down to get details about events of interest • Patterns of events?
  3. STempo - Use Where, what, when (Peuquet 2002) • overview

    of spatial patterns • exploration through filtering and other user- UI interactions • Interactive display
  4. STempo - Event Data • Current data set: data harvested

    from an RSS news feed for Syria from 2013 • High level event type categories were created: Armed group, Civil, Diplomatic • Location of the news article: georeferenced using www.geotxt.org • News article: Title, news story, source (url), date, etc. • Aggregated by event category type and geographic coordinates for the map
  5. STempo - Example Use Case Military/Tactical: We are about to

    have a US Naval Port call in Yellowland. What are the events that occurred recently and what events are likely to occur while the vessel is there, given those recent events?
  6. STempo - Pattern analysis - Political event - occurs with

    - Armed Group event - occurs with - Military event -
  7. STempo - Technology • Java application • Multiple Views of

    the data in the UI • Cross view coordination of user interactions • Statistical analysis of events generating sequences is written in Python and the results are fed to the UI • Map component is based on JMapViewer
  8. STempo UI - Before Every event location is one point

    on the map. Locations are simply overplotted. No coordination with other views.
  9. Visualization - Overview + Detail Starting with an efficient base

    map, the visualization process follows the ‘Visual Information-Seeking Mantra’ by Shneiderman (1996): Overview first, zoom and filter, then details on demand.
  10. Base map construction • Natural Earth Data • Styled with

    CartoCSS in TileMill • Exported to Map Tiles • "offline" map tiles: Java application loads map tiles extremely rapidly from those stored in the executable Java JAR file
  11. Data Visualization • What happens where, how often? • Category:

    Qualitative Data o differentiation of categories with color • Event count: Quantitative Data o only possible visual variable is size (Bertin, 1967)
  12. Current STempo UI Filter by selection: • click on one

    object • shift + click & drag to select several objects • use other components to select
  13. STempo UI example - with filtering and detail Unselected map

    symbols appear faded, matching the symbols in the other views. More detail in popup on hover of the ring.
  14. STempo Map - Major challenges • needs to be very

    flexible dealing with changing data sets • more than 10 qualitative colors for the categories are difficult to differentiate
  15. Future Work on Map symbols Visualizing temporal aspects within the

    map symbols Visualizing pattern relationships between the map symbols
  16. Future Work Contextual Geographic Information • possibility add WMS layers

    • retrieve information on climate, vegetation, population, transportation