[AI/BI] Plug and Play Data Analytics
Tl;DR
Intro
It all started from trying to talk with pandas dataframes.
And there was some evolution around it.
Chat with a DB
Data Chat RepositoryThis is not a new idea, but a new way to approach it.
And not yet trying to sell it.
About RAGs
RAGs frameworks and vector DBs have been around for few years so far.
From all the ways to do rag, Langchain is still the top 1 framework.
From the typical CSVs:
Going through PDF’s:
And most importantly: LangChain can be connected to our databases
So…how about leveling up with a cool setup?
The Setup
We will need:
- A Database to tinker with
- A way to connect the DB (inside a container) to Langchain
- A UI Wrapper to do QnA outside the terminal
- Bonus: AI/BI to get visual insights from the data
These resources will provide context:
Databases
We are going to push sample databases with tables already configured towards a container:
We can create read only users so that LangChain will connect to the DB safely.
Previously, I was using the Chinook artist related DB: via MySQL
But I want to try something new.
So it will be PGSql this time and looked with Perplexity some sample datasets.
Sample 1
Sample 2
Sample 3 - Connecting to running services
Commento
Umami
UI Wrapper
To go from a python notebook / terminal / script to something that is more non tech user friendly, we need a UI.
That’s where the power of vibe coding kicks in, together with a new project:
Why Starting a Tech Blog? 🚀
#sudo apt install gh
gh auth login
gh repo create langchain-db-ui --private --source=. --remote=origin --push
#git init && git add . && git commit -m "Initial commit: langchain x db x ui" && gh repo create langchain-db-ui --private --source=. --remote=origin --pushAs recently, I started with a BRD, some clarifications, then a development plan.
PS: You dont need 1000h of prompt engineering to do so
AI/BI
If you are kind of stucked in your D&A career, shaping one of this will be good for your portfolio.
Apache v2 | An open source alternative to Tableau. Embeddable visual analytic
Conclusions
Could this be attractive to people that have some e-commerce and dont have the bugdet to hire an BI/analyst to see whats working/whats not?
One more time, its all about the friction to PAY versus the friction to DO.
The Related Tech Talk
To unify my workflow/efforts, I’ve moved my tech talk creation from:
git clone --depth 1 --single-branch -b logtojseauth https://github.com/JAlcocerT/slidev-editor #just current status
#git clone https://github.com/JAlcocerT/slidev-editor
#git branch -a
#git checkout -b logtojseauth mainAs part of my consulting repository: the responsible for consulting.jalcocertech.com
git clone https://github.com/JAlcocerT/selfhosted-landing
cd y2026-tech-talks/langchain-postgres
#npm run dev This time I used not only components and public images, but also ./pages to keep the content modular and potentially, re-use it in the future.
Yea, thats private :)
But you can get it done for you:
Consulting Services
DIY via ebooksNext Steps
Offer Configuration
The launch strategy: aka, focus strategy
| Element | Decision |
|---|---|
| One Avatar | |
| One Product | |
| One Channel |
The Tier of Service: DIY (1b - leverages on actual tech stack Ive put together - PaaS x (WP/Ghost or SSG+CMS))
The Tech Stack:
| Requirement | Specification | Clarification / Decision |
|---|---|---|
| Frontend Framework | ||
| Styling/UI Library | ||
| Backend/Database | ||
| Authentication |
Whats Working Whats not Whats next
KPIs