Our submission for the Edelman Data Challenge 2019 contains our code used to clean, process, visualise and model the data. Our approach utilizes 2 main methods: topic modelling using unsupervised learning and network diffusion theory.
We used topic modelling, specifically Latent Dirichlet Allocation, to generate the topics from the provided articles dataset. Network diffusion theory was then applied to map the relationships between tweets and articles. The script for topic modelling is included under the LDA folder, while the scripts for network analysis are included under the network analysis folder.