How to Manually Classify Text Data – Displayr Help

If you don't want to use AI-assisted text classification, as described in How to Classify Text Data, you can instead manually classify your text data.

This article describes how to go from a raw text variable...

...to a state where the text verbatim responses have been manually classified into a new variable and can be used for further analysis:

This article contains the following sections:

Create themes
Classify text responses
Use Sort By to Assist With Classification

Requirements

A Text variable containing verbatim responses. Text variables are represented by an A next to the variable in the Data Sources tree:

Method

Create themes

In the Data Sources tree select the text variable that you would like to classify.
Hover and click + > Text Categorization.
If you want to classify responses into multiple themes, select Multiple themes, or If you want to classify responses into a single theme, then select Only one theme.
Click Start.
The text categorization window will open.
The original text responses will appear on the right:
To add a new theme, right-click in the Themes pane on the left side and select Add Theme. If you'd like to add more than one them at once, click Add and enter your list of themes, clicking "enter" between each of them and click OK.
Repeat Step 7 for any new theme(s) that you'd like to include.

Classify text responses

Next, start classifying the text responses into themes.

Select a text response on the right.
Select the theme(s) on the right to classify the text response as.
1. If you selected Only one theme when starting the categorization, the text response will be automatically added to the selected theme.
2. If you selected Multiple themes when starting the categorization, click one of the following:
  1. Classify as - to add to a theme. This option is only available if the Show responses from dropdown is set to "Unclassified".
  2. Move to - remove the response from its current theme(s) and place it in the selected theme(s).
  3. Add to - add it to the selected theme(s).
OPTIONAL: If moving to a different theme, deselect the theme(s) you want to move the text response from.
OPTIONAL: Select Remove from if you would like to remove a response from a theme.
Once you are done classifying the text responses into themes, click Save.

A new variable will appear in the Data Sources tree with "Categorization" in the name. This variable can be used in tables and other outputs.

Use Sort By to Assist With Classification

You can use our Sort by algorithms to help you manually classify remaining text data into existing themes.

Alphabetically

Sort text responses alphabetically to find items that are similar to each other or help you determine which new themes to add.

Number of responses

Use the Number of responses sort option to sort the text responses in descending order. This can help you determine which themes to add as well as assign responses into themes.

Fuzzy match

Fuzzy match sorting uses keywords to find responses that are similar in the unclassified data.

In the text categorization interface, update the Sort by dropdown to Fuzzy match.
Enter a keyword or phrase that you want to find similar matches for in the Fuzzy match field.
Click Sort now.

The responses will update to show an orange bar to the left of the text. The length of the bar indicates the match level, so the longer the bar, the better the match.

In the example below, I have some text data that contains responses to what people miss about pre-pandemic life. I used the Fuzzy match algorithm to find responses that are similar to "travel":

To assign the fuzzy match results to an existing theme:

Select the response(s) from the pane on the right that you want to classify.
Select the theme on the left to classify the response.
Optional: If you selected Multiple themes when at the start of classification, click Classify as to assign the selected items to a theme.

I can scan through the results and identify responses to classify into my existing "Missing Travel" theme.

Similarity to

The Similarity to algorithm will look at responses that have been classified to a specific theme, and then use those to find similar responses in the unclassified data. The algorithm becomes smarter as more responses are classified to a theme.

Update the Sort by dropdown to Similarity to... (selecting the existing theme that you want to find similar matches for).
The responses will update to show an orange bar to the left of the text. The length of the bar indicates the match level, so the longer the bar, the better the match.
Select the desired response(s) from the panel on the right.
Select the theme on the left to classify the response.
Optional: if you selected Multiple themes at the start of classification, click Classify as to assign the selected responses to the selected theme(s).