Statistician · M.Phil Statistics, University of the Punjab
I work on the part of data analysis that doesn't make it into the demo: the messy sensor file, the missing values that aren't missing at random, the model that looks brilliant until you fix the validation split.
My M.Phil research integrated machine learning with spatial methods to model groundwater contamination across Pakistan. The finding that stayed with me, that ordinary kriging beat XGBoost until spatial features were added, at which point the ordering flipped, is a fair summary of how I think about modelling: complexity should be earned, not assumed.
Two years of e-commerce transactions turned into a dashboard a business could actually run on: revenue, margin, AOV, cohort retention, and RFM segmentation. The headline finding, that at-risk customers hold 21% of revenue against 7% from new customers, says retention spend beats acquisition.
Python · pandas · Chart.js
Pakistan Development Indicators
35 years of World Bank data, pulled live from the API, modelled in SQL, and benchmarked against regional peers. Remittances from overseas Pakistanis are 13× larger than all foreign investment combined. Bangladesh, poorer than Pakistan in 1990, overtook it in 1998.
SQL · SQLite · window functions · Python
PM2.5 across five monitoring stations, implemented twice, once in Python and once in R. Linear regression beat Random Forest, and the README explains why that's the most interesting result in the repo.
Python · R · scikit-learn · tidyverse
Kirkgate Toolshed · live
A tool-library application built end to end for a full-stack internship: React
front end, Express and SQLite behind it, JWT auth with role-based permissions,
66 tests, deployed. Three of its bugs were found by the test suite rather than
by reading the code, including an infinite render loop that a clean build said
nothing about. The ten preceding tasks
build up to it one piece at a time.
React · Node · Express · SQLite · Vitest
Analysis that nobody can run is analysis nobody uses. This is the half that puts a model behind an interface.
Languages · SQL · Python · R · JavaScript
Analysis · Statistical modelling · Machine learning · Cohort & retention analysis · Customer segmentation · Uncertainty quantification
Tools · Power BI · Excel · SPSS · Stata · QGIS
M.Phil Statistics, University of the Punjab · CGPA 3.54
BS Statistics, Lahore College for Women University · CGPA 3.38
Statistical Intern, Pakistan Bureau of Statistics · national-level survey data
Student Statistical Consultant, LISA-LCWU · advised students across departments on study design and analysis
📧 faizajabeenfuj1999@gmail.com · 📍 Lahore, Pakistan
Open to data analyst roles, freelance analysis work, and PhD opportunities in statistics and machine learning.