Try.Databricks.Sentiment-Analysis

This code is a simple example of performing sentiment analysis with Llama 3 using Databricks' Data Intelligence Platform.

Alex Lichen and Ina Felsheim wrote a blog titled Step-by-Step Guide: AI-Powered Customer Sentiment Analysis that described how to perform sentiment analysis with Databricks SQL's AI Functions: ai_analyze_sentiment, ai_classify, etc. I found Databricks customers in the retail, consumer-packaged-goods, and quick-serve-restaurant space especially interested in this feature. The article was a great vehicle to explain how to develop this use case on the platform. However, it didn't really land with managers and less technical folk. For that, I crafted this simple 5-minute demo.

In this notebook, we create a table in a sandbox database, fill it with fake reviews, and perform analysis on those reviews. It's intended to be painfully simple.

Trying without a Databricks Workspace

Consider creating a Databricks Free Edition account if you don't have a account already. Databricks Free Edition is provided at no-cost. It is perfect for folks that want to experiment with data and AI. Navigate to the Databricks Free Edition page page for guidance on hot to sign-up then read the section "Trying with a Databricks Workspace".

Trying with a Databricks Workspace

I used the platform's Git folders feature to clone the repository into my workspace.

1. Create a Personal Access Token (PAT) in your Github account.
2. Configure the integration with Github using your PAT.
3. Create a new Git folder.