10th U6+ Consortium of African Universities International Conference

The hospitality sector is one of the significant contributors to greenhouse gas emissions yet its impact on the environment remains under researched globally. Star-rated hotels in Nairobi City engage in resource-intensive operations whose carbon footprints have not been modelled systematically using data-driven approaches. The aim of this study was to determine the impact of key hotel operations on environmental carbon footprint using machine learning (ML) among star-rated hotels in Nairobi City, Kenya. A quantitative, cross-sectional study design was adopted which targeted hotels that were rated 1–5 star by the Tourism Regulatory Authority of Kenya. Using this study design, a sample of 100 hotels was drawn through stratified random sampling. The Carbon footprint estimates were derived using emission factors from the IPCC Guidelines and the Kenya grid electricity emission factor (0.4999 kg CO₂/kWh, UNFCCC CDM ASB0050-2020). Pearson correlation analysis and multiple machine learning models which included random forest, gradient boosting, linear regression, and support vector machine regression were evaluated to quantify operational contributions of hotels to carbon emissions. The initial analyses were performed using simulated dataset that was generated mainly for code development and validation of the study methodology prior to primary data collection. The simulated values were constructed based on published literature on hotel food waste, environmental management and energy waste audits in Nairobi's hospitality sector. The study is expected to yield a predictive model to support hotel managers and policymakers in adopting targeted sustainability strategies, aligned with Kenya's Vision 2030 and SDGs.