Replace Phrase

Maintain consistency in your dataset, by finding and replacing terms in text responses.

It is common for text data to include typos, different spelling and personal preferences. For example, imagine a dataset filled with the following open-ended responses to a 'Your Australian State' question - many values can be seen under this State field:

nsw
NSW
New South Wales
Queensland
qld
QLD
QL
Victoria
VIC
Vic
NT
nt
Northen teritory
SA
south australia
TAS
Tas
Tasmania
WA

However, the desired end-state is having the following unique states in a format like - [NSW, QLD, VIC, NT, SA, TAS, WA] - to be able to generate correct insights and statistics.

πŸ‘

Replace phrase workflow

allows you to unify the values under a text field in your dataset.

How to run Replace Phrase Workflow

Once you have uploaded your data, select your dataset, click on Replace Phrase under Workflows and follow the configuration wizard.

Relevance AI - Access "Replace Phrase" Workflow

Relevance AI - Access "Replace Phrase" Workflow

In this configuration wizard, you will be asked to:

Relevance AI - "Replace Phrase" workflow wizard

Relevance AI - "Replace Phrase" workflow wizard

  • Specify which text column you wish to scan for old string values
  • Type in the (old-value -> new-value) pair (e.g. nsw -> NSW)
  • Select Yes, if you wish to receive an email notification upon the workflow completion
  • Execute the workflow

When the workflow is finalized, all specified string values under the selected column are replaced with the corresponding new values.