Week 8 – Final Week of Placement

14 Aug 2026 - Sneha Dharmeche - climate risk, data science, feature engineering, NHS ERIC

Project Overview

This week is the final week of my placement, and I had the opportunity to present my project and reflect on everything I have worked on over the past few weeks.

Just as a quick summary, my project looked at NHS estate vulnerability to climate extremes using the ERIC dataset.

I have been investigating whether estate characteristics such as building age, size and climate-related factors such as flooding and overheating can help explain/predict backlog maintenance costs across NHS sites in England.

I’ve used two machine learning models (Random Forest and XGBoost) to identify predictive patterns, alongside Bayesian regression modelling in R using brms to estimate the effects of different factors and quantify uncertainty.

Overall, the results suggest that estate characteristics, like building age and size, are stronger predictors of backlog costs, whereas the climate-related variables showed weaker associations, potentially due to the low frequency of recorded flood and overheating events.

Reflection

Looking back on the placement, one of the things I have learned is that in data science a large part of the process involves understanding the data, trying different approaches and being willing to change direction when something does not work as expected!

I also developed my confidence in working independently with datasets. There were definitely points where I found the analysis challenging, especially when dealing with missing data, skewed costs and understanding Bayesian statistics. However… working through these challenges has given me experience for future data science and bioinformatics projects.

As this is my final blog post, I would like to say a huge thank you to everyone who supported me throughout my placement.