Yeah, I think this is something we can add, but we'll need to add some pretty big caveats if we do it. Part of the problem with miles/day is that there can be gaps in the fuelling record. People can forget to add fuel-ups for years, and then come back to fuelly and pick up again. That would throw off the miles/day statistic wildly.
So we might need to find a new way to calculate "consecutive fuel-ups" or something and base that statistic on those fuel-ups only. Which will be confusing for people and harder for us to explain.
It's something we can think further about, but in the meantime you might try exporting your fuel-ups to Excel and see if you can calculate your miles/day there.
Maybe that in itself will encourage people to be more accurate in their reporting. Why try to fudge it and make it work with inconsistent data, instead of having it work with what is reported?
You already track data on an as-reported basis. For example, look at my 1994 Z. I only drive it during the warmer months, so it sits all winter. But the way the site tracks it with regards to fuel entries, it only shows the months I actually put fuel in the car. So by that math alone, you should be able to also do a daily, using the data gathered on those days/months recorded.
Use a simple mileage total for each month divided by the actual number of days that month. It will give you an average of miles per day for that month. This will change as you add fill ups of course during the month, and/or this could be a rolling month, showing your mileage average as it increases/decreases on each fill up.
Use the sum of all recorded mileage. Divide that by the number of REPORTED months. Then you can break that down by an average of 30 days. Though it will be slightly more accurate to use the previous example, this will provide a more realistic average over the ownership of the vehicle since registered and while actively updated. Either way should eliminate gap problems though.
Just an idea, you can run with it how you want. lol
Otherwise, you would have to change your entire recording structure, to count all days, not just the ones that are reported on, and figure averages that way. If that was done, your numbers would change EVERY day, fill ups added or not, because it would be adding and calculating constantly as the days go by.
My Excel sheet does this already, it uses a "Today" entry to give it the current date and uses logic (and lots of math) to count the days from a "start" date to calculate how much time to figure averages like mileage per day, hour, week, month, year, etc. I can use one fill up and know what the sheet predicts for a yearly average.
Granted, that would be a tad overkill for here. lol
But basic info, like a miles per day (and have it be for the entire garage) would be useful.
That still sounds far too complicated. There are too many rules and exceptions to blanket allow for that kind of data (some winter months, other summer months, partial months, etc). The only solution, (and I think that pb is moving to that) is that you'd have to imply something about the data - either have a "Continuous" checkbox when you report (ie this is continuous data, with no gaps) to validate that the current info should be included in the miles/day calculation, or something more complex. The problem with complexity is that people will start to not actually use it, if it is too complex.
Also, how do you handle a "long trip" - say I drive 800 miles in a day. My car can't go 800 miles on a single tank of gas, so that might wind up being 3 fillups for that 800 mile trek. How would/could you handle that on an as reported basis? I suppose you could do something like record your fillup time as well as date (which I've started doing via aCar, but that's just me), but that would be even more of a "burden" on the end user.
The reason why this site is so AWESOME is that I have reference data that thousands of other people have also contributed to. Making it "harder" (or at least more complex) to use drives away more people, thereby reducing the usefulness of the site.
So, perhaps now that I think about it, it might be useful to have that statistic with a 'caveat' asterisk.
V-bulletin calculates "posts per day" based on post-count dev into days since registration. Sure, it's not filtered by "days user didn't log in" but still represents the posts per day. Just not the number of posts per day the user was logg-in.
Missing a fuel up shouldn't be relevant. Use the all-time mileage tracked during, divided by number of days in the sample period selected: Days, weeks, months, year etc that all the mini stat fields currently have.
You hit it when you mentioned that the User is responsible for accurate data-point entry, if he/she wishes to have accurate output.
I would like to see "Average miles/day" as a mini-statistic - this would be interesting to see as it relates to how far your insurance co thinks you travel.
Originally Posted by pb
YPart of the problem with miles/day is that there can be gaps in the fuelling record. People can forget to add fuel-ups for years, and then come back to fuelly and pick up again. That would throw off the miles/day statistic wildly.
I think PB is confusing "average miles driven per day" with "average daily miles per gal/KM.
Calculating miles driven per day has nothing to do with how much fuel you put in the vehicle, nor how many times a day you fuel it. Have the routine look for the odometer for the first fuel-up prior to the beginning of the calculated period. If the vehicle hasn't been fueled for 3 years, but the period requested is 6 months, then the number of days used for the calculation is going to be 1275. If the user is reporting in Odometer readings, then it doesn't matter how many days are between the reported odometer readings.
The 'problem' with this computation will be with users that report in "miles since last fuel-up" because all the mileage intervals have to be summed before they can be divided by number of days.
Lots of mini stats are influenced by the way a User uses their vehicle. The User will develop techniques for parsing data from their vehicle's stats.
I drive my fuelly truck perhaps 3,000 miles in 10 days. Then park it for 3 months, and do it again. My "average fuel ups per week" = one, although the number of times its actually fueled in a given week is either 10 or none. So it depends on User interpretation to make the stats useful to them or not.
I too would like this mini stat. Average miles driven per day for x period.