Citrus

[mfader]

There are two weird thing in Citrus: -one very small point (=bubble) without a cross. -Croatia, which is almost on the line, appears blue, which is a strong overestimation. kind of unlogical...

[sinanshi]

can you send me your plot by email, just citrus.

[mfader]

[mfader]

samurai version:

[mfader]

also interesting: general lack of blue points in the plots: results are like this or is there a problem with the plotting of blue dots?

[sinanshi]

About the blue dot, on the graph is also pretty logical. According to the graph, LPJ value is around 2 and fao is less than 1. But I will double check on that still.

[sinanshi]

I can check the value of Crotia, however I don't know how verify the issue of lacking of blue. Do you have any idea?

[mfader]

about Croatia, makes sense I guess. I just "discover" that this way the smaller the values, the more sensible reacts the script for showing outliers.

[sinanshi]

em, what I can do is to put another two lines, one is the over estimate line, one is the underestimate line, just for testing, I can do that now, which is the easiest way

[mfader]

you mean printing to the screen the actual overestimated values? sounds good

[sinanshi]

No, you will see, I update you the graph in a min.

[mfader]

I am just thinking that maybe the problem is that the red and blue plotting is not symetric: 200% versus 50%

[sinanshi]

You are right, it should be 150% instead.

[mfader]

right, I can test myself what is the putcome of that!

[mfader]

well, delivers no change except for one dot in pulses, apparently we just do not overestimate yields a lot :-)

[mfader]

I also checked the codes in the plotting of blue dots, I cannot see any bugs, I guess it is just the results are like that

[sinanshi]

[mfader]

very interesting!

[sinanshi]

interestingly, I got more dots. :-)

[mfader]

yes! and you lost crosses (e.g. in date palm) (mmmhhhh....scratching head)

[sinanshi]

ah, what I have only changed it the duplicated green points line

[sinanshi]

well, I will sort it out, this is our last task!

[mfader]

which one did you let in in the green dots plotting?

[sinanshi]

The second one.

[mfader]

two things: -I think in the first line you plot everybody which is not zero. That is the right one. The second one only plottet when one is bigger than the other. Or I am wrong? -Second thing: Some if functions in R evaluate only the first element in an array, not elementwise. This is the difference between the functions if() else() and ifelse(), and also, the difference between using the connector && and &, if I am not remembering wrong. I am guessing the bug of the black dots is in one of these things. And that could also explain maybe the lost crosses...

[sinanshi]

I'm looking at these, you are right!! I'm a bit slow, give me some moment.

[sinanshi]

No, the first line is correct and the second line is completely stupid! To plot green cross while LPJ yields are larger than FAO. What is that?| It's only stupidity! I have removed the second line, and there is no problem. The black dot can only be made by symbol.

circle.radius <- lpj.area/max(lpj.areas)*max(fao.yields)

if fao.area is zero, and lpj.area is not zero, then circle.radius is not zero too. That is the problem!

[mfader]

ok for the green crosses plotting.

for the black dots: I don't understand what you say. FAO areas are not in the equation... And aso, I don't understand why you divide by 5

[mfader]

and also, parenthesis are a really cool thing when mixing division and multiplication ;-)

[sinanshi]

ask yourself :-D

[sinanshi]

Look at your own code, the previous Christoph's code.

[mfader]

ah ja, I remeber, they were too huge if not. In our defense, we did use parenthesis :-D

[sinanshi]

Very good.

[sinanshi]

They are solved in the latest version! Can you have a check and close this issue in due course. Thanks