Pandas profiling is a great tool for analyzing a raw data table, but often results in conflicting data types as compared to what is explicitly declared in the EML metadata for the given data table (see https://dex.edirepository.org/https%3A%2F%2Fpasta-d.lternet.edu%2Fpackage%2Fdata%2Feml%2Fedi%2F1%2F1%2Fcba4645e845957d015008e7bccf4f902 as an example). For this reason, pandas profiling should be removed completely from Dex and be replaced with a profile that is rooted in the table description as declared within the EML metadata. A simple head/tail view and describe of numeric columns would also be useful. In addition, perhaps a declared column attribute name and found column attribute view would be helpful.
Corrections to Pandas Profiling integration has now resulted in a much better user experience. I recommend keeping the current integration until demonstrated otherwise.
Pandas profiling is a great tool for analyzing a raw data table, but often results in conflicting data types as compared to what is explicitly declared in the EML metadata for the given data table (see https://dex.edirepository.org/https%3A%2F%2Fpasta-d.lternet.edu%2Fpackage%2Fdata%2Feml%2Fedi%2F1%2F1%2Fcba4645e845957d015008e7bccf4f902 as an example). For this reason, pandas profiling should be removed completely from Dex and be replaced with a profile that is rooted in the table description as declared within the EML metadata. A simple head/tail view and describe of numeric columns would also be useful. In addition, perhaps a declared column attribute name and found column attribute view would be helpful.