Funny stats from my town

Your percentages total to well over 100.

No they don't, and it's only a subset of vehicles (there's no Toyota, Honda, Ford, etc.) - the leading sentence says, the first number is the count of vehciles and the second is the percentage: i.e. SUBARU B9 TRIBECA 1 car 0.00086497707810743% SUBARU G3X JUSTY 4 cars 0.00345990831242972% SUBARU LEGACY 47 cars 0.04.6539226710492% etc.

So the total percentage comes to just under 0.116% (if those are true precentage numbers rather than fractions which then need multiplying by

But like most statistical tables, it's misleading and error-ridden. The Legacy is there three times and "Subaru" is rather meaningless, plus there's no mention of model years.

The same happened with New Zealand published "Most Stolen" lists. For a number of years the Nissan 200sx and Nissan Sylvia both appeared in the list, depsite being the same actual car.

No, after -> there is a counter of cars (one) and a percentage of that type over total number of cars of the "county." I.e. if you whant to know how many total cars there are (not listed) you have to compute

As you could see, having a suburu is a extrange phenomenom.

I didn't realize those absurd fractions were percentages and didn't look closely enough to see the % signs.

There weren't any % signs in the original post, but it did say there were three columns: "car model, a counter, and the percentage of total cars".

El 28/05/14 08:35, Your Name escribió:

I think John said he didn't realize the "absurd fractions" because he doesn't realize that here, in Europe, the market is very big and there are a lot of models, hence the market share of a good model is a success if it reaches the 1%. Then, the numbers are so low.

It doesn't help that the list is inaccurately created - it had "Subaru" as a model and "Legacy" about three different times.

El 28/05/14 09:14, Your Name escribió:

Well, the burocracy is to fill in the fields however. People doesn't know many cars models and do it as good as he can. That's a common problem, and if you think it could be solved with menus with forced choice, they would choose the wrong type.

Almost all statistical reports, surveys, polls, studies, etc. are inaccurate for a variety of reasons. The real problems come when fools like news services try and report such statistics as though they are actual facts, and can be downright dangerous when it's health reports (hence the reason apples or whatever "wonder food" are good for you one day and bad for you the next - thanks to two separate silly "studies" and their inaccuarte and misreported results).

I called them absurd because they run to as many as 15 significant digits.

You can never have enough significant digits ... just ask Pi. ;-)

El 29/05/14 02:35, John Varela escribió:

Yeah, that offending calculator machines are ruling the world.