← Back to The IndexFinal ReviewGroup 12
Final Review

Review Nº 12

Predicting Data Centre Locations Across European NUTS3 Regions
AuthorsGroup 12
24/30
Score
Clear and compact report with both classification and regression tasks, but it is too compressed and some evaluation choices make model comparisons less convincing.
Data Centres · Final Review

The Pros

1 Item
+
Addresses both presence and concentration; clearly describes dataset domains and target skew; removes target leakage columns; uses stratified split and train-only preprocessing; compares interpretable and predictive models; reports classification and regression metrics; identifies GDP and IXP infrastructure consistently; acknowledges spatial independence and missing policy variables.

The Cons

1 Item
−
The report is only three pages and lacks detail; mean imputation for all variables is weak for skewed data; top-8 feature selection by target correlation may risk leakage if not computed only on training data; Random Forest classifier does not clearly outperform the interpretable model; regression comparison mixes test-set and cross-validated results; appendix is useful but not enough to reproduce the analysis.
↩ Back to The Index
Final Review · Group 12The IndexData Centres · 2026