005 : Time Series Forecasting of Electricity Prices of a District
In this project, i explored and build time series forecasting model for measurements of electric power consumption in one household with a one-minute sampling rate over a period of almost 4 years.
About Dataset
This data set was provided by Hamoye where i did my internship as a test of our time-series analysis skill. the data contains 92016 entries and 5 columns which consists of FullDate, Tmax, SysLoad, GasPrice and ElecPrice as column names. the dataset captures the power consumption of a building for a period of 11 years (2010 - 2021)
Model Evaluation
The model's performance was evaluated using three metrics: Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Root Mean Squared Error (RMSE).
The evaluations of the FBProphet model are as follows: MAE: 70.817 MAPE: 223.174% RMSE: 91.507.
Model Analysis
Based on the evaluation metrics, the FBProphet model performed moderately well in forecasting the Electric Price of the building. The MAE value of 70.817 indicates that on average, the model's predictions were off by around 71 units. The MAPE value of 223.174% suggests that the model's predictions were, on average, 223% off from the actual sales figures. Finally, the RMSE value of 91.507 indicates that there was a root mean squared error of 92 units in the model's predictions.
Conclusion
While the fbprophet model provided a good starting point for the project, the evaluation results show that it may not be accurate enough for reliable predictions. However, there is still room for improvement. The model's evaluations suggest that it may be overestimating or underestimating the Electric prices. In order to improve the model's accuracy, further exploration and experimentation with different algorithms, models, and parameters is recommended.