Blog Archives

INSC 530 – Information Access and Retrieval

3 Credit Hours

Information access, retrieval, and use. Information seeking, user interfaces, information services and tools. Database structure, search engines, query logic, and evaluation of retrieval system performance.
Registration Restriction(s): Master of Science – Information Sciences major. Minimum student level – graduate.


Blacklight UI for PostgreSQL and Solr – Google Group Discussion

A discussion thread I started in January, 2014:!topic/blacklight-development/Ax0840hBeUA


I’ve poked around a bit and did not see anything about support for geospatial search.
I became interested in Blacklight after visiting the Environment Australia Web site.
Their site is built on PostgreSQL relational database, and since it pertains to the environment, I thought they might have spatially explicit data.
They do have “Search by Region,” but it is just a linked list of regions.
Thanks for any tips,

I got a few responses asking for more information and potential use cases so I added (in July 2014):

My opinion is any geographic representation of a collection constitutes value-added.

I am interested building a functional SKOS ontology based on the relationships between EPA ecological regions and protected areas of the United States. There are a few hierarchical levels which the ontology could describe. There are also relationships between protected areas and each ecological region.

For example, Great Smoky Mountains National Park on the North Carolina and Tennessee belongs to two main ecological regions: Ridge and Valley (67) ,and Blue Ridge (66). A full accounting of Level III ecological regions is available at <>.

The Level III ecological regions can be further divided into Level IV ecoregions, a higher level of granularity, available at <>.

Using the Smokies again as the exemplar, new ecological sub-regions at the Level IV resolution emerge: High mountains (66i), Southern sedimentary ridges (66e), Limestome Valleys and Coves (66f), Southern Metasedimentary mountains (66g), Broad basins (66j), and Southern Dissected Ridges and Knobs (67i), Southern sandstone ridges (67h), Southern Shale Valley (67g).

These regions can be linked via an ontology to other protected areas, allowing environmental information resources to be grouped based on meaningful ecological relationships. For example, a search in an information retrieval system for “High Mountains (66i)” would retrieve results from relevant protected areas: Roane Mountain State Park, Pisgah National Forest, Cherokee National Forest, AND Great Smoky Mountains National Park, and perhaps any location built in to the ontology as a member of the given ecological region’s footprint. This represents a sophisticated query with minimal effort on behalf of the user.

Along with providing a framework for information retrieval via “regions” as the Australia site does, the ontology would have useful text mining and automated spatial metadata creation applications.

Potential use cases include:

I wanted to do a Masters thesis on the impact of an ontology on search and retrieval, but after discussing with my thesis coordinator who indicated this was more of a PhD level undertaking, I opted to instead pursue comprehensive exams as my exit strategy.

I still think this is a worthwhile area of research and I am happy to see the example you have shared, and that there are others interested in the topic.

Presentation – Geocoding in Geographic Information Retrieval Systems

I presented this paper at the Geographic Information Systems II (GIS II) session at the 2014 Geography Symposium (See UT Geography Symposium Program 2014)

I represented The University of Tennessee School of Information Sciences at this interdisciplinary conference themed “Mapping outside the lines: Geography as a nexus for interdisciplinary and collaborative research.”

Tanner Jessel, School of Information Sciences, University of Tennessee. “Geocoding in Geographic Information Retrieval Systems.”

Information with a geographic component is among the most valuable and sought after types of information. However, the majority of geographical information exists as indirectly referenced locational information within unstructured text. Even among well-annotated, spatially explicit datasets, existing metadata can be of sparse, inconsistent, or otherwise of poor quality due to time and budgetary constraints. For these reasons, automated annotation of spatially explicit coordinates, a process known as geocoding, is an active area of research in geographic infor- mation science. Research concerning geocoding represents a long-term effort with a body of knowledge that has grown across several decades. Unfortunately, funding cycles are not always long-term, and some groundbreaking technologies and tools are no longer available. The present article attempts to synthesize the current state-of-the art of geocoding and presents a “toolkit” of resources used across the literature to accomplish geocoding, with an emphasis on applications for geographic information retrieval.