Data Analytics for Social Housing -

Advanced User Workshop · Two half-days · R or Python · For experienced analysts

A hands-on workshop for analysts ready to move from describing data to predicting outcomes. Working on their own organisation's data, participants explore relationships between variables and build, test and explain predictive models in R or Python.

What's covered

  • Bivariate analysis: correlation, cross-tabulation, comparing groups and testing whether relationships are real

  • Multivariate analysis: multiple regression, confounding, interactions and multicollinearity

  • Predictive modelling: logistic regression, decision trees and random forests for outcomes such as arrears, no-access or complaints

  • Model evaluation: training and test data, cross-validation, accuracy, ROC curves and avoiding overfitting

  • Responsible use: explaining models to non-specialists, checking for bias, and the ethics of predictions that affect residents

Who it's for

Analysts in performance, insight or data teams who are confident with descriptive analysis and want to build predictive models for their organisation. Participants should have completed the Introductory and Sampling courses or have equivalent experience.

Requirements

  • A laptop with R and RStudio, or Python with Jupyter or VS Code

  • Working knowledge of R or Python: importing data, basic data manipulation and plotting

  • An anonymised extract of your own data, agreed with the trainer before the workshop

Format

Two afternoon sessions of 3½ hours each (13:00–16:30), delivered online or in person. The first session builds core techniques on a shared dataset; the second applies them to participants' own data, with time between sessions to prepare.

This course is part of the Musgrave Analytics training programme for social housing, which also includes Introductory Analytics, Time Series Analysis, an Advanced Power BI Workshop, and a Manager Data Skills course.