This research has been in the works for a long time, and it feels almost strange to be finally submitting it. I have grown increasingly convinced that advances in computer science have given us new toolsets, of great value, which can and should be applied to social science questions. It is my hope that this paper takes that conviction and bestows it on the reader, and that more researchers are encouraged to explore this area. When examined using these tools and methods, this data can, and will, provide key insights into the networks of human interaction which form the societies we live in. As a field within computer science, sentiment analysis has barely existed for a decade, and it has been a great privilege to see this underutilised methodological tool used to draw findings of great social science relevance.
Firstly, I would like to thank my supervisor, Matthew Bennett, for working with me throughout this process, for his words of encouragement at just the right moments, and his critical eye which has challenged and developed the thinking behind this dissertation. Secondly, a massive thank you to Charles Rahal for providing the POTUS2016 dataset, and for the assistance in getting this project up and running. I’m grateful to Mark Lee for taking the time to sit and have an initial conversation with me about sentiment analysis, and for pointing me towards Python’s NLTK. A further major ‘thank you’ to @ItsAll_Geek2Me, and CPC for sponsoring the Raspberry PI kit which has enabled me to keep exploring the world of Twitter data mining, and given me the means to continue research in this area. Finally, I would like to thank my parents. It has meant the world to have you scrutinising my typos, dissecting my sentences and making yourselves available to discuss and develop the thinking behind this paper. Thank you.?