Week 2 – Data Exploration and Preparation

26 Jun 2026 - Sneha Dharmeche - climate risk, Preprocessing, healthcare

This week focused on moving from understanding my project aims to beginning practical data analysis work.

My current focus is on exploring NHS Estates Returns Information Collection (ERIC) data to investigate factors associated with operational vulnerability and costs across NHS sites, in relation to extreme weather events such as flooding and overheating.

Data Exploration and Preparation

A major part of the week involved becoming familiar with the ERIC dataset and understanding the structure of the available variables. Since the dataset contains a mixture of categorical, spatial and numerical information, I spent time exploring which variables may be most relevant for modelling.

I began selecting an initial subset of variables for analysis and started assessing data quality, focusing on missing values and inconsistencies in how they are recorded across the dataset. This included distinguishing between true zero values and empty or “Not Applicable” entries, as these represent different underlying meanings within the dataset and require different handling approaches.

To support this process, I began using OpenRefine to organise and clean selected variables. This involved identifying missing data patterns and considering how different approaches to handling missing values, such as imputation or exclusion, may affect later modelling stages.

---

Challenges and Reflections

One challenge was balancing project work with the heatwave this week, which made concentrating more difficult than expected.

Although I have not yet started building models, this week helped establish the foundations for the next stages of the project by improving my understanding of the data, refining the scope of my analysis and building familiarity with the tools and workflows that will support future work.