Historical Social Networks from Unstructured Data
Today we are all connected through social media. The messages we leave behind on sites like Facebook and Twitter leave a digital trail that tells the story of our experiences, preferences and relationships. Researchers can use these traces to, for example, determine public opinion or study other trends in our society.
Social historians try to do the same, but in a historical context. So without social media. Yet they would like to see how ideas spread, expand and ultimately provide an ideological change within a community. To achieve this, they often have to search through many sources and write down everything manually.
For my PhD, I therefore investigated whether and how it is possible to derive social networks from historical sources. The source I used for this is the Biographical Dictionary of Socialism and the Workers’ Movement (BWSA). The BWSA tells the story of the rise of Dutch socialism on the basis of almost 600 biographies.
Through techniques such as Named Entity Recognition, clustering and network analysis, I managed to extract a reliable network of people, organizations and locations in an automated manner. The network shows who the main players were and, when plotted over time, shows the impact of real-world events on the relationships.
|Date of defense||March 2, 2016|