Build topic models

See the example notebooks to see how to build topic models per community-detected submodules and save all variables to respective object properties.

Currently, to help name and qualitatively code the community module data, you can use the machine-learning clustering method and/or the intertopic distance method to model each module's tweet data.

Kmeans clustering method visualization

By using this method, you will also create labeled clusters for each period's modules, which you can combine and output, as desired, for further qualitative coding.

Image of example Jupyter notebook output of module's tweet data modeled via the kmeans clustering method.

Intertopic-distance method visualization

Image of example Jupyter notebook output with an intertopic distance map interactive visualization of a topic model.