I designed a dynamic Sales Analysis Dashboard for Google Merchandise Store, utilising Looker Studio to visualise Google Analytics sales data. This dashboard provided a detailed breakdown of key sales metrics, including total revenue, product performance, conversion rates, and average order value. It offered insights into customer purchase behaviours, top-selling products, seasonal trends, and revenue growth opportunities.
As my first BI tool design, this dashboard focused on sales performance by analysing revenue and sales trends.
DATA VISUALISATIONS PROJECTS
AXA UK
CUSTOMER INSIGHTS DASHBOARD
I developed a comprehensive Customer Insights Dashboard for AXA UK, leveraging the power of Databricks for data aggregation and Power BI for visualisation. The dashboard analysed over 10 million rows of customer data, delivering critical insights across 40+ demographic dimensions, including Age, Gender, and more. Key performance metrics (KPIs) such as Live Policies, MTAs (Mid-Term Adjustments), Renewals, Cancellations, and Claims were seamlessly integrated, empowering AXA’s stakeholders to make informed, data-driven decisions.
The result? A highly intuitive and scalable solution that offers deep customer insights at a glance, enabling more targeted strategies and operational efficiency.
*Dashboard cannot be shown due to confidential reasons
DATA VISUALISATION PROJECTS
AXA UK
ACCOUNT SPACE DASHBOARD
I built an interactive Account Space Dashboard using Looker Studio, transforming Google Analytics data into useful insights. This dashboard tracked the full customer journey after visiting the “My Account” page, highlighting key touch points and user behaviours. By analysing engagement metrics such as page views, AWP, GWP, journey paths, and more the dashboard provided a clear view of customer interactions and pain points.
This solution empowered the team to optimise the user experience, identify friction points, and enhance customer retention through data-driven decisions.
*Dashboard cannot be shown due to confidential reasons
DATA VISUALISATION PROJECTS
MANY OTHERS
POWERFUL OTHER DASHBOARDS
Beyond my work on major dashboards, I’ve created a variety of impactful data visualisation projects for AXA, freelance clients, and Body Jewellery Ltd. For AXA, I developed multiple dashboards including AXA’s Home Page Journey Analysis, Affiliate Transactions Tracking, Decibel Session Reporting, and journey analyses for both the Contact and Accessibility Pages.
As a freelancer, I delivered customised solutions such as Search Console Reporting, Sales Analysis Dashboards, and GA4 Report Creation for clients seeking actionable insights from their data.
In the e-commerce sector, I worked with a body jewellery ltd, building GA4 reports, Power BI dashboards for in-depth sales analysis, and conducting Market Basket Analysis to uncover key purchasing patterns.
This project is a comprehensive exploration of ensemble based machine learning techniques applied to predict hard drive failures using the massive 2022 Backblaze dataset, containing over 6 million samples. Leveraging the power of Apache Spark and Microsoft Azure for handling large-scale data, we evaluated multiple models, including Random Forest and Balanced Random Forest, across different feature selection and sampling techniques. The study delves into the effects of class imbalance and uncovers the critical role of sampling methods in boosting prediction accuracy, achieving a G-mean of over 91%. This project highlights the potential of advanced ML algorithms to optimise failure prediction and enhance hardware reliability in data-driven environments. Link to GitHub Repo
DATA SCIENCE PROJECTS
BOURNEMOUTH UNIVERSITY
FACE MASK DETECTION
This project focuses on developing an efficient face mask detection model using deep learning. Trained on a diverse dataset from Kaggle, the model achieves 90-95% accuracy with minimal computational cost, making it suitable for real-world applications. It detects various mask types (surgical, N95, reusable) across different backgrounds, angles, and facial features. The model is designed for adaptability and lightweight performance, with future improvements aimed at handling challenging conditions like facial hair and image noise. This project highlights my expertise in computer vision and practical AI solutions.
DATA SCIENCE PROJECTS
BOURNEMOUTH UNIVERSITY
HEART FAILURE PREDICTION
This project focuses on predicting heart failure risk using machine learning algorithms. Trained on the Framingham Heart Study dataset, the model analyses key clinical and demographic factors to classify patients as high or low risk for cardiovascular disease. Various algorithms, including logistic regression, decision trees, and random forests, were tested, with logistic regression delivering the best accuracy. The project demonstrates effective feature selection, model evaluation, and performance optimisation for healthcare applications, showcasing my ability to leverage machine learning in solving real-world problems in predictive analytics and medical diagnosis.
