• Small text Medium text Large text
  • Default colours Black text on white background Yellow text on black background

Using Textual Descriptions to Improve Access to Geo-Referenced Statistical Data

Principal researcher

Name: Dr. Yaji Sripada

Contact details: University of Aberdeen, Department of Computing Science
Tel: +44 1224 272597; Fax: +44 1227 273422
Email: yaji.sripada@abdn.ac.uk

Website: www.csd.abdn.ac.uk/people/homepage.php?userid=ssripada

Project details

Start date: 15/01/2007
End date: 14/01/2009

Description: We all require data or information to make decisions. For example, in order to make a decision regarding a new job to be taken up in a new area, we may like to know the difference in the average expenses between the new area and the current area. A lot of data of public interest (such as data on average expenses) vary from area to area and such data are known as geo-referenced data. The census data is a good example of geo-referenced data. In the UK, the census data tell us how various parameters such as population density and general health vary from area to area. Normally geo-referenced data are plotted on a geographical map because a map helps us to perceive the distribution of the data across the different areas of the nation. However, the data displayed using maps are not accessible to the visually impaired users because the existing screen readers cannot read out maps. A screen reader is a software tool that reads out only the textual information displayed on a computer screen. The main aim of the proposed research work is to develop techniques to automatically generate textual descriptions of geo-referenced data (map data) which can be read out to the visually impaired people using the existing screen readers. The first task is to analyse the geo-referenced data to compute how the data change across different areas. For example, given the population densities from the 2001 census for all areas of the UK, we want to discover the areas that belong to the same neighbourhood and have similar population densities. The second task is to select and organize the information related to such spatial trends into a coherent textual description. We will work with the visually impaired people affiliated to the Grampian Society for the Blind to acquire the required system building knowledge and also to evaluate our work. Although the visually impaired people will be the primary beneficiaries of our work, we believe that the textual descriptions of geo-referenced data can be useful to the sighted users as well.

Other organisations involved in this project

Last updated: 20/03/2010