This software can be used to find correlations between datasets, even if the data are relatively obscure.
For examples of the different functionality correlation provides, see the sample notebook.
correlation makes use of a few existing libraries, which you must install yourself. (numpy, scipy, matplotlib, scikit-learn)
A large portion of the analysis in correlation takes advantage of the DataFrame object in pandas.
Since numpy is a dependency of pandas, you must also install it.
pip install numpy
python3 -m pip install --upgrade pandas
The visual analysis in correlation is built on top of the powerful matplotlib.
python -m pip install -U matplotlib
Finally, the machine learning library scikit-learn which is used to perform some data analysis.
pip install -U scikit-learn
A PyPI package is in the works, but in the meantime, you mush clone the repository and use the API as follows:
git clone https://github.com/boldenth/correlation
python3
>>> import sys
>>> import os
>>> sys.path.append(os.getcwd() + '/correlation')
>>> import correlation as co
This software is licensed under the MIT License.