Scatter plotting of one variable against another (e.g., daily or monthly fuel use vs. average temperature or degree-days) is one of the mainstays of heating/cooling energy analysis. One of its limitations - finding the day/date when apparent anomalies occurred - has been overcome through the use of an Excel™ macro.
Heating/cooling energy use is generally proportional to the difference between indoor and outdoor temperatures. To describe this relationship, spreadsheets such as Excel will calculate a "best fit" equation and display a linear curve (i.e., a "trendline") conforming to it. Figure 1 shows the relationship between daily fuel use and average outdoor temperature for a two-month period. Note that many points are far from the trendline. Which indicate a problem, and which are due to acceptable factors, such as weekend closures or holidays?