Calculating fuel economy

Some recent work that Ted has been doing about his fuel economy got me thinking about my own car (a 2005 Toyota Corolla). I have been recording every fuel purchase for the life of my car, including current odometer readings, price, gallons bought, and the mileage since the last fill up. So I have all the data. I even have some ideas about trends in my fuel mileage. I just haven’t sat down and analyzed it in any form. Today, however, I punched in all the data to OpenOffice and plotted some charts to see what’s going on.

The results since winter 2004 are shown on the following graph:

First let me explain what you’re looking at. The numbers along the bottom are months. Just ignore the slashes. Each point plotted is a fill up. The blue line is simply connecting the dots between all the fill ups. The red line is a five tank moving average, which is why it starts a little to the right on the graph. Here are the trends I see when looking at the moving average, since that is a little more telling:

• There are two distinct humps in the graph - large upswings followed shortly by large falls. These humps occur early in 2005 and early 2006. This is easily explained because this is when I have taken long Christmas break road trips, which includes a lot of highway driving.

• There are two long slow upwards trends starting in the spring and going on through early fall, followed by a large fall. I believe this is a seasonal trend of getting better mileage during the warmer months with a sudden downturn when the weather gets cold and the car has to adjust.

• If you look at the tops of the peaks and the bottoms of the troughs, you will see that both are trending upwards over time. Perhaps this is explained by the car getting broken in over a couple years. I won’t know about this trend until a few more years of data are collected.

• The overall trend is upwards year-over-year. I calculated a linear regression line for the raw data (not the moving average) and found it to be y = 0.06x + 28.86. So, there is a definite upwards trend if even a slight one.

I’m not sure of what sort of driving habits might be causing these trends. This year I have gone on a lot more long trips due to hiking, which may explain why summer 2006 is slightly higher than summer 2005. In the winter, I tend to take fewer long trips and more very short trips because of the road conditions. That might explain some of the sudden plummet in mileage. However, I have not investigated my driving habits enough to full explain it. Any ideas?

It also looks like I had better get used to another downturn in fuel economy in about a month, if previous patterns continue to hold.