/jupytercon_2018

Notebooks and code of my 2018 JupyterCon talk

Primary LanguageJupyter Notebook

JupyterCon 2018

Notebooks and code of my 2018 JupyterCon talk

Dependencies

The notebooks and code depend on the following python packages:

  • numpy (Version 1.14.0)
  • scipy (Version 1.0.0)
  • pandas (Version 0.22.0)
  • sklearn (Version 0.19.1)
  • tensorflow(Version 1.2.1)
  • keras (Version 2.0.6)
  • ipywidgets (Version 7.1.1)
  • bqplot (Version 0.11.1)

Notes

  • All the examples in the talk were run in a black-themed notebook (and black-themed ipywidgets and bqplot). The colors used in the visualizations and plots were chosen to work well on the black background. Since the classic notebooks are white-themed, please update the colors (especially lighter colors like yellow) accordingly.
  • Special add-ons like full-screen button etc. were used to render the visualizations on a full screen. These are not available by default in the classic Jupyter notebook.

(Hopefully with Jupyterlab black-themed notebooks/widgets and full-screen rendering will be available for all)

The following models/algorithms/tools were covered in the talk:

  • Kernel Regression alt text

  • K-Means Clustering alt text

  • Dimensionality Reduction (PCA vs Autoencoder) alt text alt text

  • Neural Network Builder alt text alt text alt text

TODO

  • Improve documentation
  • Make notebooks run in binder