Who Likes Data?

Nice! Thanks for sharing.

 

I love data and yours is incredible! Well done! That must be be very satisfying to have. Over $7000 savings in gas! Wow!

 

Nice. You get many more miles per tank than we do. Other days we only got 5 gals in the tank from empty. Are you running the tanks past the blinking pip? You have better MPG than us which attribute to (1) looks like you try hard. (2) warmer climate. I can't tell if you are in reformulated gasoline area (as Dallas city is) but if not that might help your MPG too.

http://www.api.org/~/media/Files/Policy/Fuels-and-Renewables/2016-Oct-RFS/US-Fuel-Requirements/US-Gasoline-Requirements-Map.pdf

 

This is really interesting stuff! My husband and I have been curious about the relationship of MPG vs. ambient temperature. Our 1.5 years of data from our Gen 3 looks a lot like Paul's (blue) 2004 data here - a wide cloud of observations (lots of variation). It "feels" like there should be a relationship. Neither linear nor second-degree polynomial predictors yielded a significant model using our data.

Realizing that we had moved up the learning curve after buying our Prius in Feb 2016, I excluded the first six months of ownership. Now the model is significant. It does look like we increased our Prius-driving skills over time! I know that we aren't controlling for lots of other variables here, but 90%+ of my tanks are me driving back and forth to work on the same route (same driver, same AC preferences, same route, mostly the same traffic conditions).

Also, we don't have any observations over 90 F. I wonder if we would see a similar inflection point to Paul's (around 70-75 F) if we had more high-temperature observations.



[For reference, the equation is MPG = -0.002*Ambient(F)^2 + 0.294*Ambient(F)+ 36.9
Given our narrow range of observed temps, a linear model is only slightly more biased at high temps. The linear model (which is easier to interpret) is MPG = 0.076*Ambient(F) + 43.8. In other words, an increase or decrease in temp by 10 degrees --> in a predicted increase or decrease of 0.8 MPG.
Neither model is useful for prediction outside of the observed temp range.]

 

I live in a very hot climate, and I can confirm with anecdotal evidence that there is a *sharp* decline in MPG when it gets to be much over 100*F. Somewhere between 70 and 80 seems to be the sweet spot, based on my observations, and the data shown on this thread seem to confirm that suspicion.

 

Agreed. It's not as hot here as in AZ, but it's plenty hot and I take a hit in the summer. While the battery is admittedly outside it's best temperature when I start out, the A/C cools it down, so I suspect that much of the hit comes from the A/C itself working extra hard.