Correlation graphs are a great way to visualize and understand the relationship between features of tabular datasets.
This package leverages NetworkX, Scipy and HoloViews to easily draw correlation graphs on any Pandas DataFrame.
Works on numerical columns only for now.
This project uses Python >= 3.6.
pip install corregraphe
Example on the Titanic dataset:
import pandas as pd
from corregraphe import CorrelationGraph
df = pd.read_csv("https://raw.githubusercontent.com/mwaskom/seaborn-data/master/titanic.csv")
cg = CorrelationGraph(df)
cg.draw()Drawing gives this Bokeh figure:
- checked with
mypyandflake8 - formatted with
blackandisort - basic functionality is tested with
doctestby runningpython corregraphe/core.py
This project is licensed under the MIT License.