Hi Jason,

Is there a way to archive the entries in the mileage database, so that the people that stopped entering information (6 months or more?) could be purged?

Keep their numbers in the calculations, but archive their individual stats. Maybe give a link to access the old stuff in case someone wanted to review it?

I know you've heard it before, but Coolio (in the Insight section) hitting the #1 spot and never updating afterwards makes me wonder if that MPG is sustainable. And was that imperial gallons? To me, Excel is #1.

Regards,

 

maybe weight it. Buy avereging in a 0 for a month a user doesn't put data in. Don't use this in any calculations except displaying the chart.

 

The problem with getting rid of data is twofold: you're decreasing sample size/increasing error and you're throwing out good information. Data may be old-ish, but there's nothing wrong with it.

Imperial gallons are converted to US gallons in the database, so that isn't an issue. I could print the "last updated" date on the tables, but that could make our information look out-of-date and unusable. It isn't. I even tested taking out all cars with less than 3000 miles and seeing if the overall mileage figure changed; it didn't for any vehicle by more than 1 MPG.

I think the main issue is that fuel economy is viewed as a sort of "contest," which is exactly what it should NOT be. The second members start competing to do better, GreenHybrid looses its credibility as a reliable source of information. It will be dismissed as a sample of only the "cream of the crop" hypermilers.

 

Jason,

I think most of us now know that driving habits, trip length, and ambient temperature make up most of the differences seen in MPG.

Since trip length and ambient temperature are part of the information you collect, I think it would be interesting and useful to collate our aggregate data with these points in mind, and see it graphically.

Perhaps you could have two controls -- maybe sliders -- that let one pick a temperature range and trip length, giving bar graphs of the data subset for each car. The sliders need not be continuous; discreet groupings would work well, and would allow non-real time processing of the data.

I very much agree with your comments above, except to say that MPG is *very* much a contest. Just not against others, but a personal adventure :-)

 

Perhaps I can make the graphs even more effective, but you do realize there are already mileage/trip time and mileage/temperature graphs, right?

 

Errr .... nope. I'll go look --- hmmm. Where is it ?

Still, being able to look at a data subset that crosses trip time with ambient temp would be neat.

 

Yes, I know. I'll look into how to do that, but it's a little harder than you'd think. They're on different scales.

It's amazing... the more I do, there more to be done!

 

Jason,
Where is the trip and temp breakdown ? I looked all over the compare and learn sections. It must be staring me in the face.

Your site is wonderful. I sincerely hope it is an enjoyable hobby, and not a job ($$ considerations aside).

cheers, -- Eric

 

Eric,
I went into this as a hobby, and it's still a hobby. The graphs are the ones for the cars' detailed page. The line graph. Go to the dropdown right under it and select another graph.

 

Jason,
I think I understand your comment. I'm a pretty bad programmer, so take the following fwiw:

Rather than trying to make a single graph with two scales, let the DB build up a repository of graphs that are subsets of temperature and trip. So for example, all tanks that are 15 minute trips at 30 degrees would be one graph, broken down into bars that represent each car type. The user picks a data subset, and gets the corresponding graph.

Hope that helps explain what I am thinking about.