Personal Profile
Data scientist with experience in end‑to‑end data engineering, analysis, and modelling. I combine strong mathematical foundations with practical experience working across diverse stakeholder groups, translating complex requirements into clear, dependable data solutions. The work shown here highlights key projects across these areas.
Robust & Tailored Data Engineering
Automatic Pipelines​
When a direct API or server connection isn’t available, I use Power Automate to build secure, automated data‑collection pipelines that work within strict security constraints.
Adapted to Source
Experienced in ingesting data from APIs, SQL servers, SharePoint files, and web‑based sources (both extracts and on‑screen data), with solutions tailored to the structure and constraints of each system.
Responsible Use
Work with sensitive medical data is carried out using a disciplined, governance‑led approach. All information is handled in line with legal, ethical, and organisational requirements.
I build reliable, maintainable data pipelines that pull information from multiple sources, clean it, and combine it into well‑structured datasets. A key part of my process is understanding the precise definitions, assumptions, and limitations of the data, ensuring that any analysis or modelling built on top of it leads to conclusions we can genuinely trust.
​
​
Combining Different Data Sources
This report brings together data extracted from three GP systems — Care Connect, Anima, and EMIS. After extraction, the data is standardised and modelled in Power BI, where tables are cleaned, merged, and linked through well‑defined relationships. Adjustable targets allow performance measures to be tailored to each team or reporting period. The report is currently being transitioned to Power Automate to streamline and automate the extraction process.
Handling Different Source Data
This report uses a direct API connection to retrieve up‑to‑date market data. The data is then transformed and reformatted in Power BI, to ensure it is clean, structured, and ready for clear visualisation.
Data Cleaning
The raw data for this analysis contained a number of issues that needed resolving before meaningful conclusions could be drawn. Cleaning steps included handling missing values, correcting inconsistencies, and fixing errors using a combination of automated methods and targeted manual checks.
Clear & Informative Data Analysis
I use Power BI reports and dashboards to bring data together in one place, using interactive visuals to highlight trends and patterns. Each dashboard is designed with non‑technical, time‑pressed users in mind, making insights easy to understand and act on.
Robust and Live
Dashboards refresh automatically with new data and adapt seamlessly to structural changes, such as updated team counts or revised targets.
​​
Tailored and Insight Driven
Stakeholder priorities drive the design, ensuring visuals clearly show patterns and performance on the metrics that matter.
​
End-to-end Implementation
Led the introduction of Power BI across the organisation, including:
-
Creating staff training videos on how to access Power BI and use the full functionality of the reports
-
Delivering admin training on maintaining data pipelines and adjusting organisational targets
-
Producing troubleshooting guides to help admins identify and resolve common issues when modifying dashboards
Powerful Data Science
With a Mathematics degree from a top‑10 UK university, I combine strong theoretical grounding with practical experience in complex analysis, modelling, and predictive techniques.
Analytical Techniques
I work with simulation, semantic analysis, and numerical methods in R and Python. A strong mathematical foundation supports rigorous reasoning about structure, optimisation, and method behaviour.​
Strong Theoretical Grounding
Studying calculus, numerical linear algebra, probability, and statistics has provided a deep theoretical foundation. This enables me to understand and explain advanced techniques, evaluate their assumptions and limitations, and judge how they behave in real‑world settings.
Machine Learning
I have experience with both supervised and unsupervised machine learning, supported by a solid understanding of neural networks and the mathematical principles behind their training dynamics and generalisation behaviour.
The example classification mapping illustrates the mathematical reasoning behind decision boundaries, though it becomes computationally expensive for large n and isn’t intended as a production method.
The map on the left is taken from my analysis of grey squirrel distribution across the UK. I transformed the sighting‑location dataset by spatially aggregating records to county level and calculating relative frequencies to produce the choropleth visual.