Analytics Assistant LLM - GenAI App

Prototype. Connected to AWS and Azure MySQL DataWarehouse Build an Analytics Assistant App: Flowise AI, Text-to-SQL, FastAPI. Connect & Analyze. Any Data Warehouse. AWS-Azure MySQL

1. Detailed Implementation Guide : Friend Link (free access) to Medium.com Post

Includes step by step guide and detailed explanations.
https://medium.com/@amarharolikar/build-an-analytics-assistant-app-flowise-ai-text-to-sql-fastapi-01d8378243b4

2. YouTube Video: Walkthrough of Implementation Steps.

Build an Analytics Assistant App: Flowise AI, Text-to-SQL, FastAPI (https://www.youtube.com/watch?v=b-PcMF0Vo74)

3. LLM Assistant Video & Posts

In my earlier videos and posts, I showed how the Analytics Assistant App works… connecting to and analyzing data on AWS RDS MySQL and Azure MySQL, with table sizes ranging from 1 million to 10 million records… smoothly converting natural language text to SQL… summarizing, merging, and creating customer profiles.

This is similar to a setup I am currently using for a live client use case. This is a barebones version for sharing publicly.

Links Below:

MEDIUM: LLM Analytics Assistant: Simplifying Data Transformation & Insights. AWS & Azure MySQL LINKEDIN: GenAI App | LLM Analytics Assistant: Simplifying Data Transformation & Insights. AWS & Azure MySQL DW Example

I used a FastAPI server in the middle to allow ease of connectivity to any backend database as well as the front end of our choice.

I’ve deployed this setup on my public website (tigzig.com) using Flowise AI, and I also have a custom GPT version. In both cases, my FastAPI server and backend databases are the same.

4. What is covered in the Guide.

1. Example of Text-to-SQL
2. Implementation: Step 1 — FastAPI Server — Code
3. Implementation: Step 2 — FastAPI Server — Deploy
4. Implementation: Step 3 — Flowise AI : Setup Custom Tool
5. Implementation: Step 4 : Flowise AI : Setup Chatflow
6. Implementation: Step 5 : Flowise AI : Deploy on your Website
7. Data Warehouse — Azure — AWS — Others
8. Additional Consideration: Security, Monitoring, Tracing, Feedback
9. Cost Considerations
10. Full Data Ingestion & Split Processing: Scenarios and Costs
11. Agentic Framework, Accuracy & Complex Workflows
12. Working with LLM / GPTs for coding: My Top 6 Favorite Techniques
13. Resources

5 Flowise ChatFlow and Tool Templates

See the repository. I have exported the JSON template for ChatFlow as well as the tool agent. You can download the json files to your local drive and then upload/ import them directly to Flowise.

https://github.com/amararun/aws-azure-llm-assistant/blob/main/ANALYTICS_ASSISTANT_LLM_APP_CHATFLOW.json https://github.com/amararun/aws-azure-llm-assistant/blob/main/ANALYTICS_ASSISTANT_LLM_APP_CUSTOM_TOOL.json