GeminiBI

• 

  GeminiBI Web

Inspiration

Data Science can be very hard for most of the people and it prevents people from anlysis their data when they make decision.

What it does

This project is to help people who are not familiar with any data science topic to utilize the power of Google Gemini to analyze their data and see the connection and cohersion behind the numbers

How we built it

The project uses following tech stacks to build an completely Full Stack solution of the GeminiBI and also makes the future plan of how this solution will evolve in the future.

• Front-End: React, Apache Echarts
• Back-End: Java, Springboot, MySQL, Gemini
• Tools: EasyExcel, Mybatis Plus, Swagger, Knife4j

Backend AI Gemini (Google AI)

Challenges we ran into

• Gemini doesn't have official Java SDK support
• Can't use Gemini Model and it leads me to use the Chat
• Gemini Prompt Tunning: Sometimes Gemini ingnores my command
• Frontend utilizes the Gemini response
• API Latency

Accomplishments that we're proud of

A Complete Full Stack solution of GeminiBI with future plan

What we learned

Generally, I learned how AI can shape the world into a better place in the nearly future.

Tech:

• Gemini Prompt & how to tune it
• Apache Echarts

What's next for GeminiBI

• Java SDK support
• Gemini sometimes ignore commands

Built With

• apache
• java
• javascript
• mybatis
• mysql
• react
• springboot
• swagger
• typescript