ISSUE #001the amazing data developer · est. 2006

YASH PRAJAPATI

Friendly neighbourhood AI developer, data analyst & data engineer.

Swinging between SQL, Python & AI to turn messy data into web apps that actually help.

with great data…

THWIP!

scroll or swing

The origin story

01 / 05

Origin story, page 1: The nobody (web suit) — Page 1 · The nobodyEvery hero starts as background character #47. Ours had a hand-me-down laptop, a chai addiction and zero superpowers. The city slept. He didn't.

Origin story, page 1: The nobody (armour suit) — Page 1 · The nobodyEvery hero starts as background character #47. Ours had a hand-me-down laptop, a chai addiction and zero superpowers. The city slept. He didn't.

Origin story, page 2: The bite (web suit) — Page 2 · The bite2:00 AM. One cursed PHP file. A bug bit him HARD. Any sane person rage-quits. He got… curious. Bad news for every bug after that.

Origin story, page 2: The bite (armour suit) — Page 2 · The bite2:00 AM. One cursed PHP file. A bug bit him HARD. Any sane person rage-quits. He got… curious. Bad news for every bug after that.

Origin story, page 3: Training arc (web suit) — Page 3 · Training arcCue the montage music. Day 1: SQL. Day 90: Python. Day 200: Flutter. Day 1000: B.Tech. My spider-sense tingles. His data-sense? Never switches off.

Origin story, page 3: Training arc (armour suit) — Page 3 · Training arcCue the montage music. Day 1: SQL. Day 90: Python. Day 200: Flutter. Day 1000: B.Tech. My spider-sense tingles. His data-sense? Never switches off.

Origin story, page 4: First villain (web suit) — Page 4 · First villainEnter THE SPREADSHEET 5,000 orders. 4,999 wrong totals. Yash webbed it up, rebuilt ₹49.72 lakh of revenue from scratch and made the data confess everything. Villain: defeated.

Origin story, page 4: First villain (armour suit) — Page 4 · First villainEnter THE SPREADSHEET 5,000 orders. 4,999 wrong totals. Yash webbed it up, rebuilt ₹49.72 lakh of revenue from scratch and made the data confess everything. Villain: defeated.

Origin story, page 5: Suit up (web suit) — Page 5 · Suit upWebs weren't enough, so he engineered a whole suit: AI agents, crypto engines, data pipelines. Still a student. Already shipping. Psst… smash the suit button at the bottom.

Origin story, page 5: Suit up (armour suit) — Page 5 · Suit upWebs weren't enough, so he engineered a whole suit: AI agents, crypto engines, data pipelines. Still a student. Already shipping. Psst… smash the suit button at the bottom.

With great data

comes great cleaning

— Uncle Ben, if he worked in analytics

My
powers

← no wait →
drag your eyes this way

Power card illustration: Data analysis (web suit)

Data analysis

I interrogate data until it confesses. SQL first, Python when it needs statistics.

SQL
Python
pandas
Excel
Stats

Power card illustration: Data engineering (web suit)

Data engineering

Pipelines that move data from where it lives to where it's useful, tested so it never lies.

DuckDB
ETL
Modelling
Star schema
Data quality

Power card illustration: AI engineering (web suit)

AI engineering

LLM copilots, multi-agent assistants and ML models that earn their spot on unseen data.

Claude & Gemini
LightGBM
scikit-learn
Prompting

Power card illustration: Web apps (web suit)

Web apps

Full-stack apps from database to deploy: login, real-time data and a UI people actually enjoy.

Next.js
TypeScript
Supabase
Vercel
Flutter

Power card illustration: BI dashboards (web suit)

BI dashboards

Dashboards that answer the business question on screen one, not screen five.

Power BI
Tableau
DAX
Storytelling

THAT'SIT.(…for now. new powers drop every week)

— field reports —

The missions

dataopsv.vercel.app

DataOps workspace overview with revenue chart and data quality score

DataOps SQL studio turning English into SQL

DataOps Anomaly Watch flagging unusual days

M-01

AI data analyst platform · live

DataOps

Drop in a CSV and five AI agents take over: they score the data quality, answer plain-English questions with real SQL, build a dashboard on their own and flag weird days before you notice them.

Spidey says: it's like having five interns who never sleep. Or complain.

Next.js
TypeScript
Supabase
Gemini
AlaSQL
Vercel

Open live site ↗

coinlens · terminal

CoinLens chart view with indicators and AI trade setup

CoinLens market overview with fear and greed gauge

M-02

AI crypto market analysis · real-time

CoinLens

A live crypto terminal on real Binance data. It streams prices over WebSocket, runs 12+ indicators, and back-tests 16 strategies with fees on data they've never seen. Only setups that survive get shown. An AI copilot and a voice assistant explain every call, and a 24/7 scanner pings Telegram.

Spidey says: my spider-sense for markets, but with back-tests. And honesty.

Next.js
Python
LightGBM
WebSocket
Claude API
Supabase

Launching on its own domain soon

blinkit nexus · analytics console

Blinkit Nexus overview with revenue and delivery KPIs

Delivery analysis: minutes early or late against the promise

Stock and spoilage: products at wastage risk

Order arrival heatmap by weekday and hour

M-03

Quick-commerce data analytics · internship

Blinkit Nexus

5,000 quick-commerce orders cleaned and questioned with 92 DuckDB SQL queries and Python stats. Found 180 of 268 products stocked past their shelf life; a new stock rule frees ₹12.4 lakh. Also proved 62% of orders arrive after the promised time.

Spidey says: ten-minute delivery? The data said… not quite.

DuckDB SQL
Python
K-Means
Forecasting
Tableau

View on GitHub ↗

KeepMeFit home with BMI, calories and water tracking

M-04

Health & fitness app · team project

KeepMeFit

A mobile app for BMI and calorie tracking, personalised diet plans and nutritionist consultations, with separate client, nutritionist and admin sides.

Spidey says: even heroes need a diet plan. Fine. I'll eat the salad.

Flutter
PHP
MySQL
REST APIs

View on GitHub ↗

Damage report

Real numbers from the missions. No filters.

₹0revenue rebuilt after the source totals failed validation

0SQL queries fired

0products caught overstocked

0orders later than promised

0strategies back-tested on unseen data

0AI agents in one app

0forecast error vs 21.4% naive

— trajectory & milestones —

Journey

June 2026 – present · internship

Data Analyst Intern

keenexpert solution pvt. ltd.

Training in software development and data-driven project work: SQL, Excel, Power BI, Jupyter and DuckDB on real business data. Shipped the Blinkit Nexus analysis end to end.