What Is AI Grouping?
AI Grouping is a feature used when analyzing open-ended responses (text). It automatically corrects typos, spacing, and capitalization differences in text data entered directly by respondents, and AI groups responses with similar meanings into a single group. This allows you to quickly convert fragmented open-ended responses into quantitative data for analysis.
When Do You Need AI Grouping?
✅ When the data is scattered due to different spellings by each respondent
"I had respondents freely enter a brand name, but it came in all different. Can we consolidate them into one? 🤔"
🧙 Here's how you can use it
Applying AI Grouping automatically detects similar spellings and groups them into a single representative group. For example, if various brand name spellings are consolidated into a single value you can analyze brand awareness rankings or preferences as accurate figures.
✅ When data cleaning is difficult due to many typos or non-standard responses
"There are too many typos and abbreviations in the open-ended responses to read and classify one by one. 🤔"
🧙 Here's how you can use it
Try selecting the AI correction level (Moderate/Loose) and running auto-grouping. Quickly separate meaningless responses that are difficult to use in analysis as 'Other,' and systematically structure only the meaningful responses to maximize analysis efficiency.
How to Use AI Grouping
📌 The AI Grouping feature is available on the Professional or Enterprise plan.
1️⃣ Select the target variable on the Recode screen.
step 1. Click [Variables] at the top of the Analytics screen, then select [Create variable > Text] from the left menu.
step 2. In the target variable dropdown menu, select a 'short text' or 'ranking text' type variable for AI Grouping.
2️⃣ Run AI Grouping.
step 1. Click the [AI Grouping] button in the center of the screen.
step 2. Select your desired classification option in the pop-up and run it.
Moderate: Recognizes minor typos or spelling differences as the same response and groups them.
Loose: Actively consolidates responses even with significant spelling differences or many typos if the overall meaning is similar.
step 3. AI corrects typos and synonyms, then automatically creates groups, and 'value labels' reflecting the characteristics of each group are automatically written.
Q. Can I change the value label names generated by AI Grouping to what I want?
A. Yes, you can. The labels suggested by AI are just a guide, and users can click on the text box directly to change them to a name suitable for the analysis report at any time.
3️⃣ Edit value labels and groups.
step 1. You can directly edit the automatically generated value labels to suit your analysis purpose.
step 2. For values with too few responses or unnecessary for analysis, select [Set as Other]; for responses like 'None/Don't know,' select [Set as None] to classify them.
step 3. If you want to exclude a specific response from a group, select the value and click the [Remove] button to manage it individually.
💡 Usage tip | After setting all value conditions, clicking [Refresh response count] lets you check the number of responses corresponding to the current grouping conditions in real time.
4️⃣ Create the new variable.
step 1. Once all grouping settings are complete, click [Create variable] in the upper right.
step 2. The created new variable is converted and saved as a 'choice' or 'multiple choice' variable depending on the nature of the original variable.
step 3. The newly created variable can be used immediately for analysis like a standard choice-type question in the 'Variable list' and in the 'Results tab,' 'Cross-tab tab,' etc. of Analytics.
Do you have a good understanding of AI Grouping?
If you still have questions after reviewing this guide, please contact us anytime via the [Customer Support icon] in the bottom right corner of your screen. Our team will do its best to help you resolve any difficulties you're experiencing.