ClusterForm: Automated categorization of form textual responses using K-means

Abstract

This research addresses the inefficiencies in analysing open-ended responses collected via platforms like Google Forms. It's easy to see patterns in structured data, but you still have to do it by hand for qualitative textual responses. More and more businesses are using feedback forms and surveys, but the lack of automation in processing text data makes it harder to get timely information and make decisions. ClusterForm is a new web-based system that uses Natural Language Processing (NLP) and the K-Means clustering algorithm to sort open-ended text responses from Google Sheets. The system uses TF-IDF vectorisation for semantic analysis, which makes it easier to group similar responses. Python was used to make the backend services, and JavaScript was used to make the frontend interface easy to use. This lets users enter Google Sheet links and get clustered outputs. We used clustering performance metrics like the Silhouette Score and the Davies-Bouldin Index to test the system. The results showed that the suggested solution greatly sped up processing and improved clustering accuracy, making it a useful tool for businesses to use to automate the analysis of qualitative data.