UT Dallas Hosts Spatial Filtering Conference

Most homeowners would agree that their property values are affected by the value of their neighbors’ homes. Geospatial scholars call this underlying factor a spatial spill-over effect. Ignoring spill-over effects can skew statistical results, leading appraisers to undervalue a property surrounded by low-priced homes, for example.

Daniel Griffith
Michael Tiefelsdorf

Spatial filtering refers to a new spatial analysis technique that takes spill-over effects into account. It essentially allows analysts to extract spatial dimensions that may lead to biased results. This new discipline is the subject of an academic workshop held at UT Dallas this week.

The idea for spatial filtering was born during a brainstorming session among Daniel Griffith, Michael Tiefelsdorf and Barry Boots in the early 1990s. They’ve been developing the methodology ever since.

Griffith and Tiefelsdorf, now both geospatial information science professors in the UT Dallas School of Economic, Political and Policy Sciences, organized the workshop to teach spatial filtering to other scholars in the spatial analysis field.

The workshop is sponsored by the National Science Foundation and has attracted scholars from around the world. Admission was competitive, with less than half of the applicants receiving an invitation to attend.

Attendees include professors, researchers from public agencies, post-doctoral fellows and advanced graduate students from the United States, Canada, France, Brazil and South Korea. They come from a variety of fields — including environmental sciences, epidemiology, economics, demography and political science — but all are engaged in spatial analysis.

UT Dallas instructors include Griffith, Tiefelsdorf and post-doctoral fellow Yongwan Chun. They will be joined by renowned experts in geography and spatial econometrics.

As the organizers, Tiefelsdorf and Griffith hope the workshop helps spread the use of spatial filtering. “Ideally, participants will take away this new methodology, disseminate it and use it in their future work,” said Griffith, an Ashbel Smith Professor of Geospatial Information Sciences.

However, Griffith and Tiefelsdorf say the methodology is still under active development, and they welcome critique from their colleagues. “Due to its flexibility, spatial filtering is still a developing field. We hope participants will add their own ideas to this method,” said Tiefelsdorf.

The GIS doctoral program, housed in the School of Economic, Political and Policy Sciences, is offered in conjunction with the Department of Geosciences and the Eric Jonsson School of Engineering and Computer Science.

For more information on spatial filtering, visit the workshop Web site.

---