Data in Everyday Life: Gas Prices

With plenty of things to keep us awake at night these days, gas prices may be the least of our concerns, until we pull up to the pump for the week’s refueling and find the shock of prices that seemed to have jumped overnight—often, they do indeed jump overnight, with daily increases of 10 to 15 cents and subsequent drops of similar amounts the following day.

Reminding ourselves about the history of gas prices may be useful as we anticipate an average price of $2.57 in 2018.

At the very least, seeing this history may preclude genuine panic as we pull up to the pump. Remember the good old days of only five or six years ago? Prices hit $3.62 in 2012.  Then again, we enjoyed a $2.14 average only two years ago. What goes down must, apparently, go up.

Because of external forces such as variation in oil suppliers, global inflation possibilities, even the possibility of oil disasters, it may be hard to predict future stability of the system. PQ Systems application specialist Ryan Hartshorn’s prediction suggests that in  general, gas prices will stay within these control limits, but due to inflation and over the course of years, there will be some outliers that should be accounted for, but aren’t detrimental to the rest of the data set. Stay tuned to see how this turns out.

PQ Systems

PQ Systems


  • Very interesting.
    Mr. Harsthorn’s prediction of some potential outliers, leads me to a question:
    Considering those potential outliers (few) are not detrimental to the data set, is it wise to keep them as part of the control limits calculation? Thanks

  • > prediction suggests that in general, gas prices will stay
    > within these control limits

    I don’t see the control limits?

    I wonder if using inflation adjusted $ value would make more sense? If we did indeed get $2.57 mean this year really that is below the mean cost (due to inflation meaning those $ in the past were more valuable). That difference isn’t likely to have a huge impact in the low inflation environment over the last 15 years, but seems a more sensible measure which could really become obvious if we get a period of years with inflation that isn’t super low (which has happened plenty of times in the past).