boxuancui / DataExplorer

Automate Data Exploration and Treatment
http://boxuancui.github.io/DataExplorer/
Other
512 stars 88 forks source link

Can create_report show labels instead of variable names for labelled data? #101

Closed NatalieSoto closed 5 years ago

boxuancui commented 5 years ago

@NatalieSoto Could you be more specific? and which part of create_report are you referring to?

NatalieSoto commented 5 years ago

Sure, I apologize if my question wasn't really informative. I meant for the exploratory data analysis (Introduce, plot missing, plot_bar, plot_hist, plot_correlation). It doesn't have to be within create_report. I work with datasets with over 500 variables with informative labels and shorter (and easier to work with) variable names. When exploring the data it would be useful to see the variable labels. I was wondering if there was any way this could be accomplished (maybe by combining the DataExplorer package with other packages like expss, or sjlabelled?)that I haven't found yet.
Thank you!

boxuancui commented 5 years ago

Are these labels unique for each variable? If so, could you rename all your columns to these labels? If the labels are tag-based, you will have to process your data before plugging into DataExplorer.

In case I misunderstand your question, it will be extremely helpful if you can provide an example, ideally with reproducible code and data. Thanks!

NatalieSoto commented 5 years ago

They are unique, I just wanted to avoid renaming my columns to lengthy names. I guess this is the best way, though. Thank you!

boxuancui commented 5 years ago

One goal of DataExplorer is to get to the insights faster. I understand data can be messy, so what I usually do is to use each of the functions iteratively to get down to a smaller subset of variables of interest. Hope that helps!

NatalieSoto commented 5 years ago

It does, thank you!