Messy data.
Governed pipelines.
AI that cites its sources.

Problem

74% of new cafés in the UK fail within five years. Site selection costs £50,000+ to get wrong and most decisions rely on gut instinct and borough-level demographics that miss how dramatically conditions vary within short distances across London.

What it does

A three-task data-driven framework across 4,835 London LSOAs at granular sub-borough level. Task one: predict café success potential by area. Task two: predict commercial rent prices. Task three: find the intersection where success potential is high and rent is below market rate — that is where you open.

Key finding

Public transport accessibility dominated the model (16.92% AHP weight), above median house price, demographics, or foot traffic proxies. The more commercially valuable output was identifying emerging neighbourhoods with medium-high success scores not visible in traditional market research — areas that conventional consultancy would miss entirely.

Scale

16,361-word dissertation covering multi-source data integration, AHP weighting methodology, geospatial ML modelling, and rent prediction — then synthesised into an interactive map deployed as part of this portfolio.

Problem

Restaurant inclusivity research usually runs on small samples or a single review platform. Dietary-need signals, allergy, religious, lifestyle, sit scattered across Google Maps, Facebook and TripAdvisor as free text, across three cities, with no shared structure.

What it does

Built a multi-source ingestion pipeline across 300,000+ reviews with automated translation, deduplication and cross-platform normalisation. Classified 57,825 dietary-preference instances into lifestyle, medical and religious segments, the same microsegmentation logic consumer-insight teams run on customer data, applied here to review text.

Key finding

Compared TextBlob against VADER sentiment scoring on the same text before trusting either one, then used N-gram extraction and LDA topic modelling to surface the dietary vocabulary underneath. Shipped as Tableau dashboards for non-technical stakeholders: the analysis only mattered once someone who had never touched the raw data could read it and decide something.

Scale

Two concurrent research-assistant positions at Greenwich's Tourism and Marketing Research Centre. A parallel qualitative study from the same period, on school nutrition, went on to win Best Paper at CHME Conference 2025.