0

I am analysing a dataset with a sample of 200 000. I am only getting a table labeled as auto for correlations.

I am running ydata profiling in Jupyter with the following command

"profile = ProfileReport(df, title = "Pandas Profiling Report") profile.to_widgets()"

enter image description here

How can I get all correlations and which correlation is this? I cannot find any information about the auto title anywhere.

kehlou
  • 1

2 Answers2

0

Hey you can activate the individual Correlations as shown below.

profile = ProfileReport(df, 
  title = "Pandas Profiling Report", 
  correlations = {
    "pearson": {"calculate": True},
    "spearman": {"calculate": True},
    "kendall": {"calculate": False},
    "cramers": {"calculate": True},
    "phi_k": {"calculate": False},
  },
) 
profile.to_widgets()
Tyler2P
  • 2,324
  • 26
  • 22
  • 31
phil_o
  • 1
0

"auto" correlations calculates the column pairwise correlation depending on the type schema:

  • numerical to numerical variable: Spearman correlation coefficient
  • categorical to categorical variable: Cramer’s V association coefficient
  • numerical to categorical: Cramer’s V association coefficient with the numerical variable discretized automatically

according to the docs https://ydata-profiling.ydata.ai/docs/master/pages/advanced_usage/available_settings.html#correlations.

Simocrep
  • 3
  • 2