My Projects
Data analytics, engineering & machine learning projects built to solve real problems and grow my skills.
Analysed TfL journey data to uncover peak travel patterns and disruption impact. Mapped demand across all 33 London boroughs using GeoPandas and built an interactive Power BI dashboard.
Pulled email and social media data via SQL Server, Actito and SproutSocial, cleaned and transformed it with Power Query, and created a Power BI dashboard. Identified the top-performing campaigns.
Completed LinkedIn Learning’s Data Cleaning Essential Training. Covered missing value imputation, deduplication, string normalisation, and type coercion on messy real-world datasets.
Designed and implemented an end-to-end Azure data pipeline using the medallion architecture to ingest batch data into Azure Data Lake, transform it with Databricks notebooks, and expose curated tables for downstream analytics and reporting.
Engineered features (family size, name title extraction), handled missing data with median/mode imputation, and trained a Random Forest classifier β achieving ~82% accuracy.
Predicted residential sale prices on the Ames Housing dataset. Applied log-transformation, one-hot encoding, and Lasso regression for feature selection. Scored top 30% on the public leaderboard.
Designed and maintained email automation journeys in Pardot β including audience segmentation, personalisation rules, and A/B tests on content, timing, and cadence to improve engagement and conversions.
Analysed on-site behaviour, funnels, and traffic sources using GA4. Managed SEO components and published blog content. Collaborated on layout and CTA improvements to increase conversions.