CASE 00MASTER FILE
Want the full picture? Here's the complete file.
Mariam Rehan · Insights Analyst · Karachi
ACCESS FILECHAPTER 00 — ENTRY POINT
MARIAMREHAN
Insights Analyst (fr)
▸ I make pretty charts and tell stories. with them
hover for the truthyou're gonna love my work
Analyst.✦Storyteller.✦Chaotic Good.✦Data Witch.✦Anime Enjoyer.✦Pattern Hunter.✦Spreadsheet Sorcerer.✦Zine Collector.✦Analyst.✦Storyteller.✦Chaotic Good.✦Data Witch.✦Anime Enjoyer.✦Pattern Hunter.✦Spreadsheet Sorcerer.✦Zine Collector.✦
SCROLL
CHAPTER 01 — CHARACTER FILE
CHARACTERFILE
RPG stats for an analyst who treats messy datasets like side quests and suspicious insights like plot twists.
RPG STAT SHEET
テレパシーCuriosity99/100
Café Hours Logged87/100
Pattern Recognition95/100
Overthinking Simple Things91/100
Tolerance for Bad Data Viz12/100
EQUIPPED:
— ANALYSEPythonSQLSPSS
— VISUALISEPower BITableauExcel
CURRENTLY: Analysing consumer behaviour @ AiSight · Karachi
847
Excel Sheets Survived
2,341
Anime Episodes Watched
3.65
CGPA (Merit Scholar)
1
Research Paper @ DDW 2026
10
Cities of Consumer Data Analysed
Class: Insight Mage / Subclass: Spreadsheet Bard
ORIGIN: Institute of Business Management, Karachi · BSc Data Science · Merit Scholarship Recipient
PROFILE PIC✦ 847 XPCHAPTER 02 — CASE FILES
THE JOURNEY
Every dataset has a story. Here are the ones I've told.
CASE 01ACTIVE
Brands were bleeding customers — but not to competitors. I found where they were actually going.
AiSight · Insights Analyst · Feb 2026–present
Brand switching decomposition across multi-city consumer panels — separating competitive loss from category-level demand shifts so commercial teams could stop solving the wrong problem.
ExcelConsumer PanelsPivot TablesDirector-level Presentations
THE INSIGHT:
Brand switching decomposition across multi-city consumer panels — separating competitive loss from category-level demand shifts so commercial teams could stop solving the wrong problem.
CASE 02PUBLISHED
A forecasting model for cancer mortality trends — submitted to a global medical conference.
Broad Peak Lab · Research Assistant · Nov 2025–Jan 2026
Built an ARIMA time-series model projecting U.S. IBD-associated cancer mortality through 2035 using CDC WONDER epidemiological data. Findings submitted to Digestive Disease Week (DDW) 2026.
PythonARIMACDC WONDEREpidemiology
THE INSIGHT:
Built an ARIMA time-series model projecting U.S. IBD-associated cancer mortality through 2035 using CDC WONDER epidemiological data. Findings submitted to Digestive Disease Week (DDW) 2026.
CASE 03
Built an AI that turns messy PDFs into structured, usable forms — at scale.
Stagger Labs · Full Stack AI Engineer Intern · Jun–Oct 2025
Engineered a PDF-to-structured-form pipeline using GPT-4o, Python and Node.js — semantic field grouping, multi-page merging, incrementally ordered output. Production-grade infrastructure, not a school project.
PythonNode.jsGPT-4oLLM Orchestration
THE INSIGHT:
Engineered a PDF-to-structured-form pipeline using GPT-4o, Python and Node.js — semantic field grouping, multi-page merging, incrementally ordered output. Production-grade infrastructure, not a school project.
CASE 04
Predicted air quality 3 days out — so communities could act before the smog arrived.
10Pearls · Data Science Intern · Dec 2024–Feb 2025
XGBoost forecasting engine with serverless MLOps — hourly weather scraping, automated retraining, CI/CD pipeline, live Streamlit dashboard. Built as an intern. Runs like a product.
XGBoostHopsworksGitHub ActionsStreamlitMLOps
THE INSIGHT:
XGBoost forecasting engine with serverless MLOps — hourly weather scraping, automated retraining, CI/CD pipeline, live Streamlit dashboard. Built as an intern. Runs like a product.
CASE 05
Pakistani citizens needed a way to report broken infrastructure. So I built one.
Personal Project · Sheher Sunta Hai
Full-stack civic issue reporting platform — interactive heatmap, bilingual UI, real geocoding with OpenStreetMap. Roads, waste, water. Built for Pakistan, by someone from Karachi who gets it.
ReactPostgreSQLLeafletTypeScriptNode.js
View Project →THE INSIGHT:
Full-stack civic issue reporting platform — interactive heatmap, bilingual UI, real geocoding with OpenStreetMap. Roads, waste, water. Built for Pakistan, by someone from Karachi who gets it.
CASE 06
Taught 'Build Your First Dashboard' to women in tech — because data literacy shouldn't have gatekeepers.
AI Datayard · Market Analyst Trainee · Dec 2024–Feb 2025
Designed and delivered a Power BI training session — accessible colour palettes, beginner-friendly datasets, visual storytelling principles. Community upskilling as a form of advocacy.
Power BIData StorytellingCommunity
THE INSIGHT:
Designed and delivered a Power BI training session — accessible colour palettes, beginner-friendly datasets, visual storytelling principles. Community upskilling as a form of advocacy.
CHAPTER 03 — DATA PLAYGROUND
ME,CHARTED
Decorative viz. Real vibes. Hover everything.
24H LOADOUT
How I spend a typical day (%)
CAFÉ ORDERS (YTD)
Field research, documented
INTEREST GRAPH
Click a node. Watch the psychic links light up.
CHAPTER 04 — SIDE QUESTS
Not all quests show up on a CV. Some of them have very good ties.
…TiedTwice.pk
ACTIVE QUESTQuest: find the perfect tie. Status: ongoing.thrifted ties & belts
I sell the sickest ties and belts you've ever seen.
Visit Shop →
People STILL ask me about this tie
Why I picked it:
SCOOBY DOOBY DOOOOOOOOOO
THE BLING OMG
Why I picked it:
Belt wala showed me this and I knew he was the one
I WANTED TO KEEP THIS FOR MYSELF
Why I picked it:
Every techie needs a tie like this
CHAPTER 05 — OPEN CHANNEL
café_order.form — 注文