Data gurus, I need your help
 

Does anyone know where to get the EPA ratings for cars in ANY database format? Comma delimited is fine.

I'm going to try to add all of EPA ratings to the site so we can track who has hypermileage status, but I can't find a usable database anywhere.

A while ago someone posted a link to the EPA's downloadable database, but it was not easily usable.

If ANYONE can help, please let me know.

This will also enable you to compare your car to other drivers of the same car.

Re: Data gurus, I need your help
 

I tried to find the same thing awhile ago with no success. If you would like, I will volunteer to help build the database manually. Maybe if we all take one make, honda, acura, nissan, toyota etc... it wouldn't be such a big deal. Or..... Why not load them as we need them. Get the data for the cars that we have in the gaslog and add more as new cars are added. The EPA data goes back to 1985 so there is a lot out there.

Dibs on Honda!
 

Dibs on Honda!

Re: Dibs on Honda!
 

Honda is the only fun one for me. Please don't make me do GM!

Fine then, dibs on Acura!
 

Fine then, dibs on Acura!

Re: Data gurus, I need your help
 

If we all do it, I doubtit would take much longer than a week or two to get it all done.

I'm going on a road trip (MPG!!!) and will be back tomorrow. We can discuss it more then.

EPA 1978 -
 

EPA 1978 - 2006

https://www.fueleconomy.gov/feg/download.shtml

i opened the 1998 db and it looked ok to me.

i think rh77 pointed me to this link originally.

Re: EPA 1978 -
 

As it turns out 1998 was the first year the xcel files started looking normal. Take a look at the 1997 file. It's scary to say the least.

I guess that helps shave a few years off of our list though.

the 97 file isn't TOO bad.
 

the 97 file isn't TOO bad. just needs some massaging. i think it could be cleaned up in a few minutes.

it's not delimited and there are some unknown fields (e.g. numeric, in front of the manufacturer and a couple of other ones). but at least there's consistent whitespace between most of the columns.

Re: the 97 file isn't TOO bad.
 

I experimented with these fields today and figured I could probably import the data based on a few constants, which would make the whole thing much easier.

That being said, I started to look at the data, and it's still messy. Even the most recent years have their problems. For example, there are 4 "civics" in 1999. Nothing is said about WHICH civic they are talking about (DX, LX, Si, hatchback, sedan, etc.)

So I think the best solution might be to let the users enter their vehicles themselves. most people know their own EPA ratings. We can slowly build the database as we need it. There is no need for us to slave over 28 years of data (often with over 700 entries for each year). It also doesn't help that each year the EPA changed it's data format.

What do you all think?