Repo containing supporting material of the webinar entitled "Detecting IoT Networking Intrusion with Plotly" on April 25, 2018

You want to run these notebooks on the DataScience.com Platform? Request a demo of the Platform here.

The notebooks shared in this repo have been developed by Aaron Kramer with contributions from Jean-Rene Gauthier at DataScience.com.

Installation instructions can be found in the environment/ folder. We highly recommend pulling a pre-built Docker image we created for this webinar.

You will find the following notebooks:

• exploratory-data-analysis.ipynb: We use Plotly's Python API to explore the KDD Cup 99 dataset.

• intrusion-classification-model-build.ipynb: Following data exploration, we develop a few models to identify potential attacks from normal connections. We interpret how the different models make their classification decisions using the model interpretation library Skater

• dash-app.py: An interactive Dash app that displays graphs and tables using the same models that were developed in the Jupyter Notebooks. Run this app with $ python dash-app.py. Learn more about Dash in the Dash User Guide.

There is also a utilities/ folder that contains a series of utility functions that are used in both Jupyter notebooks.