[R] Bar Plot with Connected Points on 1 Y-Axis

[Marc Schwartz]

You can generalize the approach by using something like:

...
  ylim = c(0, max(DF$TACC, DF$Catch) * 1.1)
...


That would allow you to use the max value of the two columns, multiplied
by a fudge factor, which you can adjust as needed. In this case,
increasing the y axis range by 10% to make room.

HTH,

Marc

on 01/07/2009 02:31 PM jimdare wrote:
> Thanks Marc, that has helped a lot.  Say, for example, in a situation where I
> can't find out the highest value, is there any way to get R to automatically
> detect this and adjust the axis accordingly?  I am planning to do this for
> many different stocks at once and dont wan't to have to define the highest
> value for each.  I could set a standard axis value based on the max values
> of all stocks, however the detail will be lost for many of the less
> exploited species.
> 
> 
> Marc Schwartz wrote:
>> on 01/06/2009 09:07 PM jimdare wrote:
>>> Hi Everyone,
>>>
>>> Have created a bar plot of the data below using the following code: 
>>> barplot(TACC,space=0,names.arg=Year).  I now want to add a series of
>>> connected points to represent the catch. I tried to do this using
>>> line(Catch) or points(Catch), however both of these commands result in
>>> each
>>> data point being aligned with the right edge of each bar.  I need them to
>>> be
>>> solid points in the centre of each bar, and for each point to be
>>> connected
>>> to its neighbour by a line.  Another issue I have is when the points
>>> exceed
>>> the values for the bar graph (e.g. in 2004 and 2005 catch>TACC) R seems
>>> to
>>> cut them off, I need the axis to be expanded so they can be seen.  I'm
>>> sure
>>> these are relatively simple problems but I am really stuck.  Thanks very
>>> much for all your help, it is much appreciated.
>>>
>>> James 
>>>
>>> DATA:
>>>
>>>   Year  Species Stock TACC Catch
>>> 1 2001    ORH    OR1   5000  4687
>>> 2 2002    ORH    OR1   6000  3215
>>> 3 2003    ORH    OR1   7000  6782
>>> 4 2004    ORH    OR1   9000 10000
>>> 5 2005    ORH    OR1   9000 12000
>> One key point to note is that barplot() returns the bar midpoints. This
>> is noted in the help for barplot(). The bars are not centered on integer
>> axis values, so you need the returned values to place additional
>> annotation in the proper location relative to the bars.
>>
>> The other thing is to set the range of the y axis using the maximum
>> value in Catch, plus some fudge, so that the plot covers both sets of
>> data and has enough room for the additional points.
>>
>> Thus, presuming that your data is in a data frame called 'DF':
>>
>> mp <- barplot(DF$TACC, space = 0, names.arg = DF$Year,
>>               ylim = c(0, 13000))
>>
>> # Now use lines() to add Catch
>> lines(mp, DF$Catch, type = "b", pch = 19)
>>
>> See ?barplot, ?lines and ?points for more information.
>>
>> HTH,
>>
>> Marc Schwartz
>>