Applied case · Energy and facilities · about 70 minutes
Energy use and the weather
A year of daily meter readings from an office tower: discover why a straight line says temperature barely matters, find the curve the residuals point to, add the working-week effect and check what the residuals still say.
The brief
The facilities team at Tower 7, Lotus Business Park (a fictional six-storey office building) must submit next year's electricity budget and decide whether to fund a chiller upgrade, which only pays off if cooling drives a large share of consumption. A consultant ran a quick correlation, found it weak, and concluded that “weather explains little of our energy use — the upgrade is a poor investment.”
The facilities manager is not convinced and asks you to re-analyse a full year of daily data (energy_use, synthetic, 2023): electricity used (kWh), mean outdoor temperature and whether the day is a working day (Sunday–Thursday).
Your job: check the data, test the consultant's conclusion, build a model that fits, and say what it implies — and what it cannot tell you — for the budget and the chiller decision.
The building and its readings are synthetic, generated for teaching.
Dataset
Office building daily energy use
365 rows · one row = one day · synthetic data
| Date | Weekday | Working day | Outdoor temperature (°C) | Energy use (kWh) |
|---|---|---|---|---|
| 2023-01-01 | Sunday | true | 9.4 | 2711.1 |
| 2023-01-02 | Monday | true | 12.3 | 2658.9 |
| 2023-01-03 | Tuesday | true | 13.3 | 2560.5 |
| 2023-01-04 | Wednesday | true | 10.2 | 2670.3 |
| 2023-01-05 | Thursday | true | 10.8 | 2549.4 |
| 2023-01-06 | Friday | false | 7.6 | 1320.4 |
Included with a Formula Quest membership
This case has 4 stages and 9 graded tasks, ending with a written report.
View Formula Quest memberships