Please use this identifier to cite or link to this item: http://hdl.handle.net/1893/36167
Appears in Collections:Computing Science and Mathematics eTheses
Title: Social media mining for veterinary epidemiological surveillance
Author(s): Munaf, Amir Samuel
Supervisor(s): Swingler, Kevin
O'Hare, Antony
Reeves, Aaron
Gunn, George
Brulisauer, Franz
Issue Date: 30-Sep-2023
Publisher: University of Stirling
Abstract: Extensive records are kept in the UK regarding large-scale farms, which include information on farm sizes, locations, disease outbreaks, and the movement of animals. This data enables a nuanced understanding of the disease risks associated with commercial farms. Unfortunately, there is a lack of documented data on small-scale farms, making it difficult to evaluate the risks linked with them, despite literature inferring that they play a crucial part in epidemiological surveillance. The primary aim of this project was to evaluate the viability of using social media data as an instrument of passive surveillance for both identifying smallholding communities and early disease detection. This includes assessing the availability and quality of sufficient data, in addition to deriving meaningful inferences about the animal health population within the United Kingdom. Through the use of numerous data science techniques, such as text classification, topic modelling, social network analysis, and spatio-temporal analysis, it was possible to gain insights into the demographics, concerns, and interactions of these communities. Offering a new perspective on disease surveillance and control for policymakers, veterinarians, and agricultural experts, social media platforms have great potential to supplement traditional surveillance, as indicated by the findings. While the research faced limitations, such as the rapidly evolving nature of social media and the specific focus on English-language platforms only, it still added valuable insights to the growing body of knowledge. With the ever-increasing integration of digital and physical domains in today’s world, this research points towards new opportunities for interdisciplinary research in data science and livestock farming. Main contributions from this work: • Digital Surveillance Mechanism: Formulated an innovative methodology for monitoring and analysing smallholder discussions, concerns and actions on the internet in niche fora. • Predictive Modelling: Machine learning models have been introduced that can classify smallholding users based on their profile descriptions, providing a valuable tool for rapid identification. • Disease Outbreak Analysis: Leveraged spatio-temporal analysis to link online discussions with real-world events, providing a potential early warning system for disease outbreaks. • Network Analysis: Unveiled the complex social dynamics of the smallholder community, pinpointing crucial nodes and pathways of information diffusion.
Type: Thesis or Dissertation
URI: http://hdl.handle.net/1893/36167

Files in This Item:
File Description SizeFormat 
SamuelMunaf_Thesis_final.pdf10.35 MBAdobe PDFView/Open



This item is protected by original copyright



Items in the Repository are protected by copyright, with all rights reserved, unless otherwise indicated.

The metadata of the records in the Repository are available under the CC0 public domain dedication: No Rights Reserved https://creativecommons.org/publicdomain/zero/1.0/

If you believe that any material held in STORRE infringes copyright, please contact library@stir.ac.uk providing details and we will remove the Work from public display in STORRE and investigate your claim.