Relabel unorganized labels ! Challenge!

Question

Hi all,

Here are some facts about the data set simplified

Has a Label column that was free tex, so many people entered slightly different things

I want to organize this mess bz sorting them into the following Labels

1. No Action

2. Client request

3. Report issue

4. Data issue

5. Operations request

Also some people might have put two labels into the same cell but I of course only want one label. The numbering starting with 1. is therefore also the Hierachy say a cell has NoAction, report issue. I want the new cell to show No Action.

It would be great if somebody could help on how to solve this problem, I know it is not the easiest.

Thanks very much!

Below find data examples

Next I will input some Label examples, that it would need to handle

reports-issue

reportissue

data-issue, reportable_ruledata-issuedata-drivenclientrequestReport-IssueReport-Issue, data-issueOperations-RequestOperations-Request, data-issue

KGT · Answer

I'm not sure what pre-prep you did before fuzzy match, but the way you would most likely want to go about this is:

* RecordID
* Text to Columns: Split to rows based on a comma delimiter
* Multi-row to give each recordID set it's own ID counting up (Group on recordID, RowID = [Row-1: RowID]+1
* Match the exact matches to your labels
* Clean up free text etc, this could involve fuzzy match to get the list of labels in each group (eg data-issue matching to Data Issue)
* Look at your leftover and work out how that would be grouped.

Fuzzy Match is usually an iterative process as you want to set the limits at a level not to get false positives.

l_v_fernandez · Answer

I used fuzzy match and it works for almost everything/ But the Hierarchy is of course not implemented and it doesn't show all labels unfortunately. Grateful for any help!