Liam Brannigan
-
Profiling a Polars query
-
ML pre-processing with Polars
-
Data science 2025
-
Polars 🤝 Matplotlib
-
Polars & Huggingface datasets
-
Polars on a diet
-
Polars up and running
-
Polars ❤️ sorted data 1: statistics
-
Polars doesn't have an index but what if you want one? Or many?
-
Fit XGBoost models directly from Polars and Arrow
-
Don't fear the List dtype in Polars
-
Breaking your bad habits with Polars
-
Getting the largest values with Polars
-
Combining data with different schemas
-
Polars can help if your data is sorted
-
Using Spacy with Polars?
-
Flexible time series aggregations in Polars
-
Doing ML pre-processing in Polars
-
Setting an attribute shortcut in python
-
Don't loop over columns in Polars
-
IPC files in Polars
-
Loading from a database with Polars
-
Boolean filters in Polars?
-
Testing for data science 3: testing with Pandas
-
Testing for data science 2: testing in practice
-
Testing for data science 1: why test?
-
Attaining flow while writing software
-
What software skills do oceanographers want to learn
-
Dev in docker 2
-
JPO at 50: Part II
-
JPO at 50: Part I
-
Dev in docker
-
Career progression
-
Bayesian workshop
-
Correlation Bayesian style
-
Fast looping with Numba
-
Bayesian inference of MOC variability
-
Dipole breakup
-
Submesoscale instabilities in mesoscale eddies
-
Sliding through the data with Bokeh
-
Easy animations in matlab
-
Python repo for MITgcm
-
And a site was born
Workshops & talks
This workshop that aims to show how effective python is for data science. This was developed originally for climate scientists, but in practice anyone can follow it.
This workshop introduces Bayesian statistics. This was developed originally for climate scientists, but in practice anyone can follow it.
Presentation at Python Ireland Meetup in March 2018 introducing Bayesian machine learning.
Visualisation
I love making animations using matplotlib! However, it can be a pain to make these animations at the start. I've put together the following module that allows you to make animations using a declarative syntax. See this github repo for the code and examples.
Interactive visualisations are an effective way of exploring your dataset. This snippet shows how to use the Holoviews library to do this in a jupyter notebook.