Louisiana, Minnesota and Dallas shooting topic modelling

This is visualisation of VU Amsterdam summer school course "Big Data in Society" final project. Data about Dallas shooting (and shootings that were previous to it) was gathered from New York Times Twitter, Facebook page and articles. After data cleaning topic modelling was used. Topic models are here:

Topics popularity over time:

Topics popularity in time

Facebook (left) and NYT (right) corpora comparison:

Facebook and NYT corpora comparison

Facebook (left) and Twitter (right) corpora comparison:

Facebook and Twitter corpora comparison

Twitter (left) and NYT (right) corpora comparison:

Twitter and NYT corpora comparison

Cosine similarity of Facebook and NYT topics:

Cosine similarity of Facebook and NYT topics

Cosine similarity of Facebook and Twitter topics:

Cosine similarity of Facebook and Twitter topics

Cosine similarity of NYT and Twitter topics:

Cosine similarity of NYT and Twitter topics

Sentiment dataset per dataset (boxplot):

Sentiment dataset per dataset

Sentiment dataset per dataset (density plot):

Sentiment dataset per dataset

Sentiment dataset per dataset and topic(boxplot):

Sentiment dataset per dataset topic

Sentiment dataset per dataset and topic(density plot):

Sentiment dataset per dataset topic

Topics network (was visually same for all datasets):

Topics network

Code for the project is here. Authors: Pablo Bernabéu and Risto Hinno.