Reports
How to use:
- Choose a forecasting model
- Select a campus to view its data
Model
Campus
Line Chart - Interactive
Showing Actual vs Predicted Electricity Consumption
Essential Details of UNICON Dataset
UNICON Building Data Distribution
Distribution of cleaned data points across different UNICON Buildings
UNICON Campus A: 858,235
UNICON Campus B: 164,160
UNICON Campus C: 47,484
Total Data Points: 1,069,879
Note: The total hourly data points are aggregated during the SARIMA modeling process. Thus, from 1,069,879 data points, it has been reduced to 52,308 data points.
Spearman Correlation Results
View the correlation matrix showing the statistical relationships between different variables used in the forecasting model.
Note: Boolean variables, including event types, were one-hot encoded before Spearman analysis. Air Temperature and Event Types were dropped due to high correlation (>|0.85|) causing redundancy and multicollinearity.
Exogenous Variables
Environmental and demographic factors used in the forecasting model
Weather Parameters
| Variables | Units |
|---|---|
| Wind Speed | meters per second (m/s) |
| Relative Humidity | percent (%) |
| Apparent Temperature | degrees Celsius (°C) |
| Air Temperature | degrees Celsius (°C) |
| Dew Point Temperature | degrees Celsius (°C) |
Other Exogenous Data
| Variable | Unit |
|---|---|
| Capacity | number of people |
| Is Holiday | boolean |
| Is Semester | boolean |
| Is Exam | boolean |
| Event Count | number of events |
| Event Type | — |
| Recent Events | number of events |
| Event or Holiday | binary |
Note: These exogenous variables provide crucial environmental and demographic context that influences electricity consumption patterns across different public school locations.