I have conducted an online survey where 1500 users participated. They have their opinions and comments about particular problems. I am trying to consolidate all their opinions & pulling out patterns from their comments. It is not easy to read through all 1500 comments and create pattern quickly for a single person.

Is there any recommended method which can help me analyze 1500 comments and draw couple of patterns out of these comments?

2 Answers 2


There are many methods, and choosing one depends upon what type of research you are doing. Specifically how precise and valid does your findings have to be.

In qualititave research grounded theory and thematic analysis are two methods used in academia. For ux-studies I would recommend using an approach based on thematic analysis. This means that there are two stages analysis and synthesis.

This article on analysis and synthesis gives a good overview of the general steps:

 1. Collect and organize the data. Make your data manageable.
 2. Mine the data. Identify what you see.
 3. Sort and cluster the data. Manipulate or reframe your data, as necessary.
 4. Identify insights. Discuss, articulate, incubate, and socialize your insights. 

There are tools, methods and strategies for doing this work. For instance using card sorting to sort comments into groups. I suggest using excel, google sheets or airtable together with some ideation tool like realtimeboard. There are other tools that I haven't used like optimal workshops reframer, however it is quite expensive.

Its also interesting to read how professional ux researcher think and work, therefore I would suggest looking at polaris nuggets for inspiration.

  • Thanks. This will surely help. But I was looking for some tool which quickly can help me analyze 1500 comments and throw some pattern or wish list of users.
    – Souvik Das
    Feb 28, 2019 at 17:28
  • 1
    My recommendation is to use excel, google sheets or airtable. Create a «row» for every new meaningful datapoint (encoding). For each comment have a unique ID and add it to the row. When another comment included that datapoint update the counter for that row and add the ID. Include quotes or special insight that relates to each row. Have another researcher do the same to create validity. Check correlation between findings by using the unique ID and see if that uncovers any patterns - user who have an issue with «a» wanted feature «b» or «c».
    – simonlh
    Feb 28, 2019 at 19:41

Ideally you should have asked the questions in a more analytics friendly way. For example, you should limit the number of open ended questions you ask and try to provide more multi-select questions. For multi-select question you can easily generate charts.

There are form creators that allow you even to take the users in a particular flow if they responded in a certain way. This allow for more detailed answers in a particular topic. Something like:

  - 1 - 2 - 2.1 ................... - 3 - 4
          - 2.2 - 2.2.1 - ......... -
                - 2.2.2 - 2.3 - 2.4 - 

Otherwise, you can try to:

  • Take all the words and try to see pattern occurrences. You can try word/tag clouds like https://www.wordclouds.com/ . Paste you data in and see which words get duplicated. enter image description here
  • After you have your most used words you can search them for the context they appear in.
  • You can also try some pre-grouping. For example, create a tag cloud just for the data from the users that provided variant A for the question 1.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.