Distance Matrix in python from pandas dataset

Question

I have the following pandas dataframe

import pandas as pd    
df = pd.DataFrame(zip(["A","B", "C", "D"],[10,30,55,60]), columns=["Name", "Distance"])

Out:

  Name  Distance
0    A        10
1    B        30
2    C        55
3    D        60

Now, I would like to make a distance matrix, i.e.

Does anyone know how to make this efficiently with python?

Look into https://scikit-learn.org/stable/modules/generated/sklearn.metrics.pairwise_distances.html — Learning is a mess, Jan 13 '22 at 10:29
Does this help ? https://stackoverflow.com/questions/64008893/create-a-distance-matrix-from-pandas-dataframe-using-a-bespoke-distance-function — sagi, Jan 13 '22 at 10:34

mozway · Answer 1 · 2022-01-13T10:42:54.500

from scipy.spatial.distance import cdist
pd.DataFrame(cdist(df[['Distance']], df[['Distance']]),
             index=df['Name'], columns=df['Name'])

from sklearn.metrics import pairwise_distances
pd.DataFrame(pairwise_distances(df[['Distance']]),
             index=df['Name'], columns=df['Name'])

or, simply, raw numpy with broadcasting:

a = df['Distance'].values
pd.DataFrame(abs(a-a[:,None]), index=df['Name'], columns=df['Name'])

NB. the scipy and sklearn approaches enable to use a wide range of distance functions, not only the euclidean distance.

output:

Name   A   B   C   D
Name                
A      0  20  45  50
B     20   0  25  30
C     45  25   0   5
D     50  30   5   0

1 Answers1