/python-science-tutorial

Series of notebooks to illustrate different plotting features using Python

Primary LanguageJupyter NotebookMIT LicenseMIT

Python Science Tutorials 📈📊

This repository will contain a series of scripts and notebooks to help people get acclimated to using Python for scientific publications.
Follow me at: @naveen.venkatesan


Table of Contents

  1. An Introduction to Making Scientific Publication Plots with Python

  2. Basic Curve Fitting of Scientific Data With Python

  3. Visualizing Three-Dimensional Data - Heatmaps, Contours, and 3D Plots with Python

  4. Intro to Dynamic Visualization with Python - Animations and Interactive Plots

  5. A Guide to Creating and Using Your Own Matplotlib Style

  6. Generate Easily Reproducible Scientific Figures with Pylustrator

  7. Create Panel Figure Layouts in Matplotlib with Gridspec

  8. Intro to Comparing and Analyzing Multiple Unevenly Spaced Time-Series Signals

  9. An Introduction to Plotly for Matplotlib Users

  10. Creating Various Plot Types and Subplots with Plotly


An Introduction to Making Scientific Publication Plots with Python

Link to Article: https://towardsdatascience.com/an-introduction-to-making-scientific-publication-plots-with-python-ea19dfa7f51e

Notebook: python-plotting-intro.ipynb


Basic Curve Fitting of Scientific Data With Python

Link to Article: https://towardsdatascience.com/basic-curve-fitting-of-scientific-data-with-python-9592244a2509

Notebook: curve-fitting-tutorial.ipynb


Visualizing Three-Dimensional Data - Heatmaps, Contours, and 3D Plots with Python

Link to Article: https://towardsdatascience.com/visualizing-three-dimensional-data-heatmaps-contours-and-3d-plots-with-python-bd718d1b42b4

Notebook: heatmaps.ipynb


Intro to Dynamic Visualization with Python - Animations and Interactive Plots

Link to Article: https://towardsdatascience.com/intro-to-dynamic-visualization-with-python-animations-and-interactive-plots-f72a7fb69245

Notebook: animation.ipynb


A Guide to Creating and Using Your Own Matplotlib Style

Link to Article: https://towardsdatascience.com/a-guide-to-creating-and-using-your-own-matplotlib-style-30de6c60bac0

Style File: scientific.mplstyle


Generate Easily Reproducible Scientific Figures with Pylustrator

Link to Article: https://towardsdatascience.com/generate-easily-reproducible-scientific-figures-with-pylustrator-9426292e07a4

Python Script: plot.py


Create Panel Figure Layouts in Matplotlib with Gridspec

Link to Article: https://towardsdatascience.com/create-panel-figure-layouts-in-matplotlib-with-gridspec-7ec79c218df0

Notebook: gridspec.ipynb


Intro to Comparing and Analyzing Multiple Unevenly Spaced Time-Series Signals

Link to Article: https://towardsdatascience.com/intro-to-comparing-and-analyzing-multiple-unevenly-spaced-time-series-signals-e46b2347972a#9512-73cf3e25e1a0

Notebook: time_series.ipynb


An Introduction to Plotly for Matplotlib Users

Link to Article: https://towardsdatascience.com/an-introduction-to-plotly-for-matplotlib-users-9f4f0d2113bc

Notebook: plotly.ipynb


Creating Various Plot Types and Subplots with Plotly

Link to Article: https://towardsdatascience.com/creating-various-plot-types-and-subplots-with-plotly-bd727f808262

Notebook: plotly-charts.ipynb