Data Visualization in Python

Matplotlib is the foundation: static, publication-quality figures with control over every element, at the cost of verbose code. Seaborn builds on it with statistical plots that take DataFrames directly and good defaults. Plotly produces interactive HTML charts, and its plotly.express API makes common charts one-liners. Altair is declarative, built on Vega-Lite, and expresses charts as a grammar of encodings; it embeds the data in the chart spec, so very large datasets need aggregation first. Bokeh targets interactive browser plots and includes a server for live apps.

To turn analysis into an app, Streamlit reruns your script from top to bottom on each interaction, which makes it the fastest way to a prototype. Dash, from the Plotly team, uses explicit callbacks and gives more control over layout and state in larger apps. Gradio is aimed at machine learning demos and makes it easy to share a model's interface. Graphviz renders graph diagrams, and Cartopy draws maps with proper projections on top of Matplotlib.