Data Exploration with Snowflake Arctic

app screenshot
data visualization screenshot

Inspiration

When handling a new dataset, it's always a struggle to come up with insights and explore the data effectively. Now, with the power of Large Language Models (LLMs), the process of data exploration has been significantly simplified.

What It Does

Our project creates a workflow for uploading a CSV dataset and utilizes LLM to generate possible prompts from the dataset, considering its schema. It then uses these insights to generate data visualizations.

Challenges We Ran Into

We spent a considerable amount of effort on prompt engineering to tune the LLM to output the most desired and structured response.

Accomplishments That We're Proud Of

The overall experience of our project exceeded our expectations when we started. It seems that LLM has a clear picture of how to navigate through different types of data, which column of data is worth exploring, and what type of graph is suitable for the exploration. We think this project will empowers people with no coding background with a strong capacity for data exploration, and knowing that makes us proud!