README.md
README.md
2. EDA, analysing the seasonal patterns and outliers.
3. Filtering the data as per the use e.g., Resampling & Removing outliers.
4. Resampling as per different rules and analysing the different patterns after plotting all resamples.
5. Checking the stationarity of data.
6. Plotting acf and pacf charts for both non seasonal and seasonal data and analysing p,d,q as well as P,D,Q.
7. Train test split our monthly resampled data.
8. Applying Seasonal ARIMA or say SARIMAX, and plotting the results.
9. Calculating R2 score and residual means, which are as 0.98 and ~2.6°C
10. Forecasting the data for next year end. Seasonal Arima is able to predict the monthly average temperature with R2 score nearly 98% which is pretty good. The error in predicted temperature on train data is around 2.6°.