Welcome to usualsuspects’s documentation!

This is simply a personal collection of out-of-the box tools for making visualisations commonly found in ML papers.

Note

This is primarily a personal project with the aim of practicing python packaging and good software engineering practices whilst developing something useful for me personally. If it helps you that’s great! However, there are much more fleshed out visualisation aides out there done by large teams and dedicated individuals.

Feel free to explore how the tools are used in the examples found in the Github repo.

Installation

Installing usualsuspects

A stable version can be downloaded from PyPI

pip install usualsuspects

Installation from sources

The most recent version can be cloned from the Github repository and then built from the project root folder

pip install .

If you wish to modify some of the source code for your own task, you can make a developer install. It is slightly slower, but immediately takes on any changes into the source code.

pip install -e .

License

MIT License

Copyright (c) 2020 Paul Scherer

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

usualsuspects package

usualsuspects.tsne module

class usualsuspects.tsne.Quick2DTSNE(X, y, perplexity, initialisation='pca', n_jobs=-1)[source]

A quick tool for creating 2D T-SNE plots with input X, and optional true classifications y for coloring.

From Scikit Docs: t-Distributed Stochastic Neighbour Embedding is a tool for visualizing high-dimensional data. It converts similarities between data points to joint probabilities and tries to minimize the KL divergence between the joint probabilities of the low dimensional embedding and the high-dimensional data. T-SNE has a cost function that is not convex. i.e. with different initializations we are bound to have different results.

This tool is a shortcut to producing the T-SNE plot one finds so often in papers.

Parameters:
  • X (array-like) – asdfasdf
  • y (array-like) – asdfasdf
  • perplexity (float) – The perplexity is a hyperparameter that is related to the number of nearest neighbours to consider. Typically larger datasets need a higher perplexity to look good.
  • initialization (str (default="pca")) – Possible values = [‘pca’, ‘random’] describes the initialization.
  • n_jobs (int (default=-1)) – The number of cores to use. Default is -1 which uses all available system cores
Returns:
Return type:

Quick2DTSNE
plot_embedding(title=None, save_to='tsne_plot.png')[source]

Creates a matplotlib pyplot and saves it as a png image.

Parameters:
  • title (str) – Title for the plot
  • save_to (str) – path and filename where the image will be saved. Extension is contextual but expects .png or .jpg
Returns:

Saves a PNG image of the plot to the path in save_to
Return type:

None

usualsuspects.pca module

class usualsuspects.pca.Quick2DPCA(X, y)[source]

A quick tool for crating 2D PCA plots with input X, and optional true classification y for colouring

From SciKit Docs: Linear dimensionality reduction using SVD of the data to project it to a lower dimensional space. The input data is centered but not scaled for each feature before applying the SVD

This tool is a shortcut to producing PCA plots

plot_embedding(title=None, save_to='pca_plot.png')[source]

Creates and saves a matplotlib pyplot of the first 2 eigenvectors of the PCA operation on X

Parameters:
  • title (str) – Optional title for the plot
  • save_to (str) – path and file_name where the image will be saved. Extension is contextual but expected behaviour is to use .png
Returns:

Saves a PNG image of the plot to the path in save_to
Return type:

None