given the following bar graph (which shows monthly revenue, but no actual values), is there a way to calculate the revenue (actual dollar amount) in 3/2013?
Asked
Active
Viewed 2,678 times
0
-
It's approximately 7.8 times greater than the value in 5/2011. – Kaster Apr 25 '13 at 08:01
-
Question revised to remove ambiguity. Thanks for pointing that out @Kaster – yas4891 Apr 25 '13 at 08:04
-
No, not without some kind of scale or at least one $y$ value. Why don't you have the $y$ values...? – icurays1 Apr 25 '13 at 08:04
-
@icurays1 It is a random graph picked off the internet, solely fore asking this question. I remember having read an article on the topic, but can not find it online anymore. IIRC the method described in the article was able to deliver a good estimate for the actual y values – yas4891 Apr 25 '13 at 08:06
-
@yas4891 As it was pointed out, there's no way to figure that out without knowing the value for any of those bars. – Kaster Apr 25 '13 at 08:11
4 Answers
1
It is impossible to determine even a rough or order of magnitude estimate. They could draw an identical graph if all their revenues were a million times larger than they were, and you would be none the wiser.
not all wrong
- 16,178
- 2
- 35
- 57
-
@TheGreatDuck Unless "Time = Money", that unfortunately doesn't make any sense. The x-axis is measured in units of days, and it is meaningless to talk about measuring currency in units of days. (If it was a scatter plot of two variables measured in currency, and not a time series graph, then you could indeed indicate scale of one axis and state the other is the same; and that would be enough - but still awful practice!) – not all wrong Mar 16 '16 at 10:47
0
Obviously there is no way of knowing the y axis, Though I would question the similarities to German Tank problem. Another way is to determine whether the thing in question (money values) increased exponentially or decreased exponentially and does our graph have this curve inside of it? This would in the long run allow us to determine if the y axis is linear or logarithmic.
Question for the comments are there other main types of axis besides linear and logarithmic? How would we check for these in the long run?
shai horowitz
- 2,086
0
There is an old fashioned pantograph method that runs on area calculation using Green's theorem.
Take maximum value as 8 units, the maximum is a bit less than 8 units. A month is one unit $ \Delta t$ .
One needle is fixed arbitrarily, another goes around the boundary and a graduated wheel measures loop area. Since $t$ increment is known, total revenue and average monthly can be calculated.
$$ \bar y = Area/ ( \Delta t .\cdot base ) $$
Narasimham
- 40,495
0
If you expressed each revenue in terms of the initial revenue (12/2010) you could model the growth as an exponential function. $$R=R_{12/2010}*e^{\gamma*t}$$ However without knowing a dollar amount for at least one of the data points I cannot think of a way to solve for the exact dollar amount in 3/2013.
user2307487
- 178