Scanning all new published packages on PyPI I know that the quality is often quite bad. I try to filter out the worst ones and list here the ones which might be worth a look, being followed or inspire you in some way.

scikit-learn-inspired time series.The primary goal of this library is to allow one to train time series prediction models using a similar API to `scikit-learn`. Consequently, similar to `scikit-learn`, this library consists of `preprocessors`, `feature_extractors`, and `pipelines`.

A safe, transparent way to share and deploy scikit-learn models. Other methods for exporting scikit-learn models require Pickle or Joblib (based on Pickle). Serializing model files with Pickle provide a simple attack vector for malicious users– they give an attacker the ability to execute arbitrary code wherever the file is deserialized. (For an example see: https://…on-pickle-security-problems-and-solutions ). sklearn-json is a safe and transparent solution for exporting scikit-learn model files.

A modern, enterprise-ready business intelligence web application

A lightweight deep learning library

Image classification using tensorflow.

A statistical package to manage data

Repair erroneous entries in data streams while maintaining accuracy of overall trends. Yayes patching was designed to repair erroneous entries in data streams while maintaining accuracy of overall trends. The yayes package is most powerful in situations where users see sudden, temporary drops in variables’ values that then return to correct variable values (i.e. discrepancies in variable values that may be caused by missing data or data input errors). This method was inspired by the logic behind max-pooling insofar as high values of data often contain more meaningful information and that lower values of data may more often be discarded without losing relevant information.

bert for tensorflow

Manage your dataflows seamlessly

A library for data science