How to Supercharge Your Python Classes with Class Methods

 How to Supercharge Your Python Classes with Class Methods | by Siavash Yasini | May, 2024 | Towards Data Science

As we just mentioned, a class method is a method that is bound to the class itself and not the instance of the class. This means it can change the state of the class such that it applies across all current or future instances of the class. A famous and practical example is when you want to create a singleton class. 

Singleton is a design pattern that allows you to constrain a class to only have one instance. Here’s an implementation:

And that’s the whole idea behind classmethods. They allow us to define methods in a class that are bound to the class itself, not its instances, therefore they allow us to modify the class behavior and make it more flexible.

pandas tricks

 a list of aggregation function for groupby:

min, max, mean, median, sum, count, std, var, size, describe, nuniaue, idxmax, idxmin.

df.groupby([‘airline’, ‘weekday’])[‘cancelled’,’diverted’].agg([‘sum’,’mean’]).head()

df.groupby([‘airline’, ‘weekday’])[‘cancelled’,’diverted’].agg({“dist”:[‘sum’,’mean’]).head()

pandas tricks

 df.style. highlight_max(axis=‘colums’)

CatBoost vs. LightGBM vs. XGBoost

 CatBoost vs. LightGBM vs. XGBoost | by Kay Jan Wong | Towards Data Science

 

What is a Generalised Additive Model?

What are Generalised Additive Models? | Towards Data Science

5.3 GLM, GAM and more | Interpretable Machine Learning (christophm.github.io)

pandas tricks

 df['count'].clip(lower=1, upper=10)

pandas tricks 3

 df.sort_values(["a","b","C", ascending=[True, False, True)

df.nlargest(10, 'a').nsmallest(5,"b)