Sentiment analysis used to be a pain. You needed complex coding skills or expensive specialized software. Not anymore. You can do it for free with tools you probably already have.
If you’re tired of reading through row after row of spreadsheet data trying to figure out how people actually feel about something, here’s how to let NotebookLM do the heavy lifting.
The Workflow
The process is straightforward. It connects tools you’re likely already using.
- Create a Google Form: Design your survey with open-ended questions. (These are the ones that give you the rich, unstructured feedback.)
- Export to Sheets: Make sure your form responses are going to a Google Sheet. This is usually the default.
- Connect to NotebookLM: Add that Google Sheet as a “Source” in a new Notebook.
Once your data is in NotebookLM, that’s where the magic happens. You can ask it to run reports, identify themes, and specifically analyze sentiment. Ask things like, “What is the overall tone of the feedback regarding X?” or “What are the top three frustrations mentioned in column B?”
Going Beyond Text: Creating Visuals
Getting a text summary is nice. Using that data to create better deliverables is better.
NotebookLM can suggest structures for infographics, slideshows, or visual presentations based on the data patterns it finds. Instead of handing your team a raw spreadsheet, you can hand them a visual story.
How I Used This in the Real World
Prompt: Create a breakdown infographic for Group 1. Use the Boise State Logo, and the standard Boise State Colors.
I recently used this exact workflow to process feedback from 10 student groups. They had completed a data analysis assignment and presented their findings.
I took the raw survey data into NotebookLM and asked it to generate a summary report specific to each group. Instead of sending everyone a generic overview, I was able to send a targeted summary to each team.
But I didn’t stop at the text. I asked NotebookLM to help me visualize the data. I used its output to create an infographic for each group that represented their specific summary.
This worked incredibly well. It transformed what would have been a boring “wall of text” email into something engaging. An artistic visual element. It made the feedback accessible, easy to digest, and much more impactful.
You can find examples of all of this inside this Google Drive Shared folder. Also, the workflow instructions are much more verbose.
What else could be done with this workflow? Here are a few ideas I have had since I created this post.
- This was created 2 days before Gemini could be used as a data source in a Gemini prompt. This post is a few days after that, and my Gemini EDU account still can’t attach a notebook. It only works in my personal Gemini account. Since I would never recommend uploading student data in your personal account, I can’t exactly test this out on this workflow yet. I don’t know if or how that would change this workflow.
- I did this as part of a final, but if you are using it in an ongoing class, you could generate a slide deck on it as well and give a more general after-assignment feedback in your next class.
- Using the new data table studio feature, you could probably look at the data in different ways and use Google Sheets to look at the data in different ways. I could imagine a number of different ways to combine the text and numbers to create differing data visualizations. Number of groups that did something. Groups that did X got Y score, stuff like that.
- Leave a comment on other ideas on where this could go.
