Binning algorithms

Binning is the process of taking a continuous variable and categorizing it into discrete buckets. It can be useful to turn a potentially infinite amount of values into a finite amount of “bins” for your analysis.

How to do it

Let’s imagine we have collected survey data from users of a system. One of the survey questions asks users for their age, producing data that looks like:

df = pd.DataFrame([
    ["Jane", 34],
    ["John", 18],
    ["Jamie", 22],
    ["Jessica", 36],
    ["Jackie", 33],
    ["Steve", 40],
    ["Sam", 30],
    ["Stephanie", 66],
    ["Sarah", 55],
    ["Aaron", 22],
    ["Erin", 28],
    ["Elsa", 37],
], columns=["name", "age"])
df = df.convert_dtypes(dtype_backend="numpy_nullable")
df.head()

        name       age
0       Jane       34
1       John       18...