When executing pandas-profiling, there was an argument that can be used in cases where you want to skip high-load processing such as correlations and dynamic binning, so make a note as a reminder.
skip_conf
profile = ProfileReport(large_dataset, minimal=True)
profile.to_file(output_file="output.html")
・ Enter ** minimal = True **
・ When you want to apply pandas-profiling to a large data set ・ When pandas-profiling causes an error in correlations or dynamic binning
https://github.com/pandas-profiling/pandas-profiling#large-datasets
Recommended Posts