Climate change data science

Of the many triumphs of human understanding, the ability to extract meaning from noisy real-world data is one of the most practically powerful. Climate data is a particularly good example of this, because the underlying question is simple and important, but the evidence has to be handled with statistical care. During my undergraduate degree, I had the privilege of working on an extensive data analysis assignment using the Met Office UK Climate Series, applying statistical and data science methods to investigate whether the dataset provides significant evidence for climate change.

This project gave me a valuable opportunity to demonstrate my ability to turn raw historical data into a coherent quantitative argument. I analysed long-term temperature, precipitation, seasonal, and frost-day records using techniques such as baseline anomaly construction, ordinary least squares regression, hypothesis testing, distributional analysis, correlation analysis, and statistical visualisation. The assignment was also a useful exercise in communicating uncertainty: not just identifying trends, but testing whether they were statistically meaningful and interpreting the limitations of the results.

I was pleased to receive 90% for this work, which made it one of the clearest demonstrations of my data science and statistical reasoning skills during my degree. The following report presents the finished analysis, showing how the UK climate record provides strong quantitative evidence for accelerated warming, changing temperature distributions, increased warm extremes, reduced frost days, and shifts in precipitation behaviour.