SOCIAL MEDIA ANALYSIS CLASSIFICATION USING MACHINE LEARNING
Awarded a Distinction for my MSc dissertation at Birmingham City University, this research applied Machine Learning and NLP to classify over 44,000 COVID-19 lockdown tweets into sentiment categories. By comparing algorithms — from Random Forest and Naïve Bayes to deep learning models like LSTM and BERT — the study measured how effectively social media can reveal public mental health trends. The work highlights how advanced analytics can support institutions in making better, data-driven decisions during crisis situations. Access the RESEARCH HERE.
Medical Research analysis THESIS
True Stats are supporting and executing a clinical research project examining the association between non-alcoholic fatty liver disease (NAFLD) and dietary habits among patients undergoing ultrasound in an emergency setting. We are leading the entire data workflow — from raw survey extraction and rigorous data cleaning to variable harmonisation, recoding, and statistical analysis planning. Complex multi-response dietary and lifestyle variables were systematically standardised into analysis-ready formats, allowing meaningful comparisons across NAFLD severity grades. The final dataset is structured to support descriptive epidemiology, association testing, and advanced modelling, enabling clear, clinically relevant insights for thesis submission and future publication.