This project was developed alongside a comprehensive literature review of 7 peer-reviewed climate science papers, identifying key gaps in existing research:
- Most existing studies rely on limited data sources — this project addresses that by analysing 250+ years of records across 243 countries
- Current research uses outdated statistical methods — this project applies ML-based predictive modelling as an improvement
- Existing models lack regional granularity — this project identifies priority regions with the highest temperature increases
- Enhanced Data Processing — built a preprocessing pipeline using two missing value strategies (removal vs. mean imputation by country) to maximise data retention while maintaining statistical validity
- Historical Trend Analysis — comprehensive time-series analysis of long-term climate behaviour from the 1850s to present
- Regional Pattern Identification — identified regions with the most significant temperature changes, useful for climate policy prioritisation
- Seasonal Variation Analysis — quantified seasonal temperature differences to support climate change adaptation models
- Predictive Modelling — linear regression forecasting framework serving as a reference for both research and policy planning
- Global average land temperatures have risen steadily since the 1850s baseline
- Identified the top 10 countries with the highest recorded temperature increases
- Mean imputation by country produced cleaner trend lines than row removal, better preserving regional climate variation
- Predictive model projects continued warming if current trends persist, consistent with the 1.5–2°C thresholds discussed in IPCC literature
This project was completed with a full literature review and academic report.
📄 View the Literature Review Report