EXPLORATION INTO DEVELOPING DAYLIGHTING ZONES FOR NIGERIA USING CLIMATE BASED DAYLIGHT MODELLING TOGETHER WITH SUPERVISED AND UNSUPERVISED STATISTICAL TOOLS.

Daylighting zones have been developed for some countries predicated on existing climatic zones. This paper assesses the Koppen – Trewatha and Horn climate classification for daylighting zones in Nigeria (via the simulations of glazed and unglazed secondary school classrooms) using CBDM (Climate-based daylight modelling) methodology and analyses with the supervised learning statistical tool of ‘Chi square goodness of fit’ test. It was discovered that there was no ‘goodness of fit’ between the maps generated by the datasets generated and the climate map of Nigeria based on the Koppen – Trewatha and Horn classification. Therefore, using the unsupervised learning statistical tool of ‘Hierarchical Clustering’ an exploratory data analysis resulted in the development of a 2-cluster ‘Daylight Zone’ for the classrooms. This study concludes that the ‘exploratory’ daylighting zones/subgroups generated by this paper provides a better climate classification and is worthy of further development. Keywords: Daylighting Zones, Koppen-Trewatha-Horn, Climate-Based-Daylight-Metrics, Supervised and Unsupervised Statistical Tools