Data Analyst · Bath, UK
MSc Data Science student at the University of Bath — I specialise in product analytics, statistical modelling, and building self-serve tools that let teams answer their own questions. I write findings clearly enough for a C-suite audience.
Product Analytics & Experimentation
End-to-end A/B testing (frequentist Z-test, Chi-square, and Bayesian). Funnel drop-off, retention cohorts (30/60/90-day), RFM segmentation, LTV/CAC. I care about whether an intervention worked, not just whether a coefficient is significant.
Analytics Engineering
Production SQL modelled with dbt (staging and mart layers), Streamlit and Power BI dashboards built for stakeholder self-service. I build the infrastructure that removes the analyst from the critical path.
Statistical & Bayesian Modelling
Hypothesis testing, regression, Bayesian hierarchical models (NumPyro/NUTS), time series (SARIMA, Prophet, LSTM). MSc modules in Applied Data Science, Statistical & Bayesian Data Science.
Executive Communication
Gap analysis, key driver analysis, structured recommendations. I translate model outputs into one-page executive summaries and video presentations for non-technical stakeholders at C-level.
Pulled 509M taxi trips, 93M subway records, and Census income data to study who gets hurt by NYC's congestion toll. Ride-hail fell 7%. Yellow cabs rose 9.5%. Traced the 58-point demand gap to transit access, not income. Modelled platform substitution (2.65×) and projected a $130M revenue swing from a proposed 2027 toll increase. Argued transit investment protects low-income riders better than fare exemptions. Published in full on Zenodo.
frequentist &
Bayesian agreed
Built retention, funnel, RFM, and LTV/CAC analyses on 31,524 events using dbt. Ran A/B tests — both frequentist and Bayesian found the same +31.6% uplift. Built a 6-page Streamlit dashboard so stakeholders stopped asking the analyst.
deployed as
live API
Benchmarked 6 models on 36,500 rows of retail data. Moving Average 7d won at MAPE 14.9%. Built an inventory optimiser (EOQ, safety stock, reorder points) and deployed it as a live Docker API with 4 endpoints.
London Grocery Consumption Analysis
Merged 1.6M Clubcard records with ONS income data. Found ready-meal consumption has no link to income. Delivered as a video + infographic to a C-level audience.
National Student Survey Feasibility Analysis
Recommended revamping Business Studies to a university executive board. Found a −2.5% teaching quality gap and built a composite satisfaction index (r = 0.93 with official metric).
Bayesian Component Degradation Modelling
Built a hierarchical Bayesian model (NumPyro/NUTS) to predict mechanical failures. Cut unexplained variation by 75%. Correctly flagged high-risk components in a blind test.
Diabetes Progression Statistical Analysis
Found BMI predicts disease progression strongest (p < 10⁻⁴¹). Built a sex classifier at 73% test accuracy using linear discriminant analysis. Delivered as a 10-page structured report.
Click2Flick — Automated Educational Video Generator
Won 1st Prize at a national 12-hour hackathon. Converts a text description into an educational video automatically using NLP + D-ID + FFmpeg.
Code Revive — Legacy Code Migration Tool
Shortlisted from 300+ teams at EG AI Hackathon 2024. Converts COBOL codebases to Python, Java, or C++ using LLaMA 3 — 100% correctness on all 3 test codebases.
I'm available for a Summer 2026 placement (ongoing MSc) and open to data analyst, product analyst, and BI analyst roles in the UK. If you have a hard question and a dataset, reach out.