plotnine VS causalml

To your last point, have you tried plotnine? It's meant to be ggplot2 for python.

https://github.com/has2k1/plotnine

plotnine - A grammar of graphics for Python based on ggplot2.

Github: https://github.com/has2k1/plotnine

This seems quite similar to plotnine [0], which also provides a grammar of graphics interface for Python. That said, I love ggplot and I can't wait to use this in my research! I hope we can port/re-implement ggthemes, scientificplots [1], and other ggplot libraries for lets-plot.

0: https://plotnine.readthedocs.io/en/stable/

1: https://github.com/garrettj403/SciencePlots

There is [plotnine](https://plotnine.readthedocs.io/en/stable/) which tries to implement ggplot in Python.

You could use plotnine if you like the grammar of graphics concept: https://plotnine.readthedocs.io/en/stable/

> Python doesn’t yet have anything remotely close to ggplot for rapidly making exploratory graphics, for example.

Plug for plotnine (https://plotnine.readthedocs.io/en/stable/). I don't know R but use ggplot indirectly through this library for exploratory data analysis, and comparing the experience to any other python plotting library, I understand why R folks are usually so sad to be using Python.

They have, https://plotnine.readthedocs.io/en/stable/

Uplift Modeling (python): CausalML, EconML

Do you (or Microsoft DS teams in general) tackle any causal problems? Apparently, uber tries to solve these issues, they created https://github.com/uber/causalml

There's some causality focused packages. Nothing like scikit-learn or statsmodels yet because it's all based on very recent research, but this one: https://github.com/uber/causalml

I agree that uplift trees and CATE methods are promising here. If you’re in Python, check out Uber’s open-source causalml.

Many libraries have recently emerged that offer implementations of algorithms for heterogeneous treatment effect estimation (or, CATE estimation). The most well-known examples are Microsoft's EconML (https://github.com/microsoft/EconML) and Uber's CausalML (https://github.com/uber/causalml). Existing libraries require all data to fit in memory, which is often a limitation for industry applications on web scale datasets. Booking.com's new library offers similar functionality on top of Spark, enabling web scale uplift modeling.

you like causal inference? it must be nice to be able to use libraires like dowhy, causal ml, and ananke right? 🤔🤔🤔

video's author recommends this course in a comment: https://www.coursera.org/learn/crash-course-in-causality and he also co-created this library: https://github.com/uber/causalml

There aren't many general solutions because it will really depend on your domain. There are some packages that specifically work for "modeling under interventions" (like this https://github.com/uber/causalml although I have never tried it). In general, if you have enough data between intervention and not-intervention, you could train two different models and then apply whichever one makes sense (e.g. if you wanted to find the highest-churn-risk users among those who have already had the intervention, use the model trained on the prior intervention cases).