DMP: Cultural Hunger: Predicting British Museum Attendance via National Bread Prices

Context and methodology

This project investigates the "Cultural Hunger" hypothesis: a spurious correlation study exploring whether the retail price of a standard 800g white loaf can predict monthly visitor attendance at the British Museum. The goal is to demonstrate a complete, FAIR-compliant machine learning pipeline, from data integration to model evaluation.

The experiment re-uses two official UK government datasets: "Museums and Galleries Monthly Visits" (DCMS) and the "Retail Price Index for Bread" (ONS). The data has been integrated via temporal alignment (matching records by month and year), cleaned for missing values using Python/pandas, and split into training (70%), validation (15%), and test (15%) subsets for supervised regression analysis.

Technical details

The project artifacts are provided in open, interoperable formats:

• Data: Tabular datasets are stored in CSV format (e.g., cultural_hunger_train.csv), containing monthly time-series features for bread prices and visitor counts.

• Model: A Linear Regression model serialized as a .pkl file (cultural_hunger_regressor_v1.pkl).

• Metrics: Model performance (MSE and R2 scores) is logged in a JSON file (model_evaluation_metrics.json).

• Visualizations: Trend histograms and scatter plots are provided in PNG format. The workflow is designed for reproducibility using Python 3.x and standard libraries such as scikit-learn and pandas.

Further details

This experiment is conducted as part of the FAIR Data Science (DaSt 2026) course. While the relationship between the variables is intentionally spurious, the methodology strictly follows FAIR principles to ensure the experiment is findable and reproducible. All source data is used under the UK Open Government Licence v3.0, and the produced artifacts are available for re-use under a CC BY 4.0 license with proper attribution to the author.