Common Python packages

Here are some of the most common Python libraries. They are open source, widely used, and well supported libraries.

Included in Anaconda

Data Science pandas - Data manipulation for tabular data (think a super Excel)
numpy - Speed and memory optimized arrays for numerical computing. Underlying library for many other libraries (pandas, scipy, scikit-learn).
scipy - Collection of math and scientific functions
statsmodel - Statistical functions.
scikit-learn - Machine Learning algorithms.
matplotlib - Low level plotting library.
seaborn - High level statistical plotting library.
bokeh - Interactive plots that are converted to JavaScript.

Web/HTTP
requests - HTTP for humans
flask - Web application framework (REST, templated html, etc.)
gunicorn - HTTP web server

Not in Anaconda

Data Science
TensorFlow - General computation library, often used for creating Neural Networks.
Keras - Higher level API for TensorFlow (makes TensorFlow easier to use).