A comprehensive analysis of Rome's historical weather patterns from 1950 to 2022, examining temperature trends, precipitation patterns, and climate change indicators through statistical analysis and machine learning approaches.
The analysis reveals significant temperature patterns in Rome:
- Clear seasonal temperature cycles
- Long-term warming trend visible in the data
- Monthly temperature distribution shows peak temperatures in July-August
- Significant temperature variations between seasons
Key precipitation findings include:
- Highest rainfall typically occurs in February (~225mm)
- Driest month is September (~5mm)
- Clear seasonal precipitation pattern
- Notable year-to-year variability in rainfall amounts
The correlation analysis shows:
- Strong positive correlation (0.99) between average and maximum temperatures
- Strong positive correlation (0.98) between average and minimum temperatures
- Weak negative correlation (-0.35) between precipitation and temperature variables
The dashboard provides:
- Long-term temperature trends
- Temperature distribution analysis
- Monthly temperature patterns with standard deviation
- Average monthly precipitation
- Temperature vs. precipitation relationships
- Special weather conditions frequency
- Model performance comparisons
- Prediction error distributions
- Time series analysis
- Seasonal decomposition
- Statistical testing (Mann-Kendall, Shapiro-Wilk)
- Machine learning models (Random Forest, Linear Regression)
- Data visualization using matplotlib, seaborn
The analysis included multiple machine learning models:
- Linear Regression (R² = 0.884)
- Ridge Regression (R² = 0.884)
- Lasso Regression (R² = 0.838)
- Random Forest (R² = 0.918)
The analysis uses the 'Roma_weather.csv' dataset, containing daily weather records from 1950 to 2022, including:
- Average temperature (TAVG)
- Maximum temperature (TMAX)
- Minimum temperature (TMIN)
- Precipitation (PRCP)