Case study

Using natural language processing to structure market research

Learn how a market research start-up used classification and clustering to gain insights to their free-text survey responses.

This guidance is part of a wider collection about using artificial intelligence (AI) in the public sector.

AI techniques and application used:

  • classification
  • clustering
  • natural language processing

Objective

A market research start-up needed to access insights from their free-text survey responses.

Situation

A market research start-up was collecting free-text survey responses rather than structured answers. This presented a challenge, as the startup needed to analyse insights in real-time. The number of responses the start-up was gathering meant this was not something a human could do.

Action

The start-up applied a classification algorithm to the text data to match survey responses to pre-identified trends. The start-up then used natural language processing to analyse over 15,000 free text survey responses to identify consumer trends relevant to a particular client.

In parallel, the start-up built a clustering algorithm to draw out themes in the data and identify unexpected topics and responses.

Impact

The startup deployed the model as a dashboard which allowed employees to gain real-time insights and rapid analysis of free-text survey responses. This is a notoriously difficult action to do given the highly qualitative nature of free-text surveys. By clustering topics and responses, the start-up could efficiently unlock valuable insights and themes to inform their ongoing business strategy.

Updates to this page

Published 10 June 2019