Time Series Analysis with Polars: Exoplanet Transit Detection

A beginner-friendly tutorial that teaches Polars (a modern, fast data manipulation library) by analyzing real NASA space telescope data to detect an exoplanet transit.

What You'll Learn

• Load and explore data using Polars' eager and lazy modes
• Write Polars expressions and chain operations together
• Understand when and why to use lazy evaluation for query optimization
• Handle missing data and create derived columns
• Perform time series operations (rolling windows, binning, grouping)
• Compare Polars and pandas performance on real tasks
• Create both static (matplotlib) and interactive (Bokeh) visualizations

What We're Analyzing

This tutorial uses real observational data from NASA's TESS (Transiting Exoplanet Survey Satellite) mission. We'll analyze a phase-folded light curve from the hot Jupiter exoplanet WASP-18 b to:

• Detect the transit signal (when the planet passes in front of its star)
• Measure the transit depth (observed brightness decrease caused by the transit)
• Calculate transit properties
• Create beautiful visualizations of the light curve

Prerequisites

You should be comfortable with:

• Python basics (functions, loops, variables)
• DataFrame concepts (rows, columns, filtering, grouping)
• Basic pandas (if you've used df[df['col'] > 5], you're good!)

You do NOT need:

• Deep pandas expertise
• Prior Polars experience
• Astronomy knowledge (we explain everything!)

Getting Started

1. Clone (or download) this repository

git clone git@github.com:dbouquin/polars_demo.git
cd polars-timeseries-tutorial

2. Create a conda environment

We recommend using conda to manage your Python environment and dependencies:

# Create a new conda environment named 'polars-tutorial' 
conda create -n polars-tutorial python=3.13 -y

# Activate the environment
conda activate polars-tutorial

3. Install required packages

Install all necessary packages using conda:

conda install -c conda-forge polars pandas numpy matplotlib bokeh pyarrow -y

4. Download the data

The tutorial uses the CSV file in this repository called wasp18b_lightcurve.csv, which contains WASP-18 b light curve data. You can use this file or build your own using fetch_data.py:

python fetch_data.py

5. Execute the notebook

Open run polars_timeseries_tutorial.ipynb using your polars_tutorial environment

Repository Structure

polars-timeseries-tutorial/
├── README.md                          # This file
├── polars_timeseries_tutorial.ipynb   # Main tutorial notebook
├── wasp18b_lightcurve.csv             # Light curve data 
└── fetch_data.py                      # Script to download data (optional)

Acknowledgements

• Built with Polars, pandas, matplotlib, Bokeh
• Data fetch and plotting adapted from STScI Archive Scientist Susan Mullally's Exoplanet Data and TESS Light Curves Using Python Requests
• Tutorial developed for the Anaconda community