What is Topic Extraction?

You can convert open-ended data into quantitative codes with Phonic's topic extraction.

Phonic is a platform that aims to bring better insights from Qualitative data. One of the ways we do this is with topics.Topics can be extracted from new data automatically, assigned manually, or you can run a custom auto-classification once data is collected.

Auto-generated Topics

You can enable automatic topic extraction in the survey builder under the Analysis section of question settings. When this option is selected, topics will be automatically extracted from any new responses according to a generic set. Topic extraction must be enabled for each question in your survey.

Manual Topic Extraction

Topics can also be generated manually using the tools within Phonic. Click on your survey and then click on your question of interest. If the topics feature is enabled, you'll have the ability to manually apply topics to each response. This can be done in addition to any automated topic extraction to enhance accuracy.

Phonic also provides some additional tools to make manual labelling easy. Merge, add, rename, and delete themes all in the "Topics" tab when in Question view.

Auto-classification

The final way to apply themes to your data in Phonic is to use auto-classification. Here, you'll have the ability to specify a list of topics under which you'd like to classify the rest of your data. Simply visit the "Topics" tab in Question view, and click the button titled "Auto-Coding". You'll be greeted with a popup asking for a list of themes. If you'd like to try out a few classifications before running on your entire dataset, we recommend sampling your data and running some tests here.

Simply click "Try a Sample" and then "Load a Response". Enter a few classes and click "Run Sample" to see some results. Once you're happy with the classes you've selected, click "Run Classification" to start the classification job.