University of Twente Student Theses
Using Natural Language Processing to interact with geospatial data
Embrechts, V. (2024) Using Natural Language Processing to interact with geospatial data.
|
PDF
1MB |
| Abstract: | Geospatial data can provide valuable insights into environmental patterns, (urban) development and resource management. Currently the analysis and interpretation of geospatial data requires knowledge of complex softwares and methods, effectively limiting the power of geospatial science to a small group of knowledgeable individuals. Background research shows a lack of current end to end natural language systems that perform question answering while referencing geospatial data. Using the Creative Technology design process and co-design, multiple natural language question answering systems were prototyped, during specification one system was chosen to be built for production. In a constant loop of evaluation and optimization individual modules were improved. The system met the requirements of being speedy, accurate and non-proprietary / self hosted. The tested speed was under 10 seconds on average, the system accurately answered questions 92% of the time, struggling more with complex questions than simple ones. Lastly the system can be run on a local machine and does not depend on proprietary software. This new geospatial question answering system allows inexperienced users to interact with and use geospatial data to improve our world. |
| Item Type: | Essay (Bachelor) |
| Clients: | SuPerWorld |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 17 linguistics and theory of literature |
| Programme: | Creative Technology BSc (50447) |
| Link to this item: | https://purl.utwente.nl/essays/101915 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page