Problems exporting ggplot2 graph as PDF using ggsave

Question

Here is my dataframe using dput():

AggBar <- structure(list(Visual = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 
    3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 
    5L, 5L, 5L, 5L, 5L, 5L), .Label = c("Left (40ms)", "Left (90ms)", 
    "SIM", "Right (40ms)", "Right (90ms)"), class = "factor"), Audio = structure(c(1L, 
    1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 
    3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 
    3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L), .Label = c("Left Beep 1st", 
    "No Beep", "Right Beep 1st"), class = "factor"), Report = structure(c(1L, 
    2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 
    3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 
    1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L), .Label = c("Left", 
    "SIM", "Right"), class = "factor"), Prob = c(0.83837774456841, 
    0.0582319772265122, 0.103390278205078, 0.376964076581909, 0.533271295345261, 
    0.0897646280728304, 0.491246357933239, 0.138301223292309, 0.370452418774452, 
    0.967653676751111, 0.00938599635932435, 0.0229603268895651, 0.877493727693669, 
    0.0720126632217432, 0.0504936090845876, 0.876575180584984, 0.00762664462699526, 
    0.115798174788021, 0.492520101540142, 0.29634748728839, 0.211132411171468, 
    0.0853528134021257, 0.803322136564915, 0.11132505003296, 0.155147359805197, 
    0.250345330474703, 0.5945073097201, 0.301905843152117, 0.140332387165982, 
    0.557761769681901, 0.0572000601130054, 0.503570857472542, 0.439229082414452, 
    0.0700895268446092, 0.0751875639266702, 0.854722909228721, 0.100018323936676, 
    0.0125390266853255, 0.887442649377999, 0.0431475160621075, 0.068269348744299, 
    0.888583135193594, 0.0235399740203181, 0.0042071142101905, 0.972252911769491
    ), SD = c(0.132514958276699, 0.0456130005972609, 0.118780003012801, 
    0.135944534822725, 0.143955013979635, 0.0570946432581132, 0.197212542689416, 
    0.107930186460654, 0.177082458822958, 0.0408127107496353, 0.0199945720490766, 
    0.0364346081328841, 0.133304906228258, 0.0819834012009533, 0.0660651710233039, 
    0.121660680076734, 0.01958675910736, 0.117439307692672, 0.205346740985586, 
    0.197451097807221, 0.190264024509737, 0.0770279870403184, 0.144412752633359, 
    0.0856446869844664, 0.201202913892, 0.161633110440843, 0.185655750581604, 
    0.138719102634961, 0.104516121191566, 0.166845454185104, 0.0580320672043022, 
    0.17918882055406, 0.155080139699953, 0.0923569889436162, 0.063020378352565, 
    0.104479039597072, 0.0856849197498157, 0.0208368690589841, 0.0967941898575448, 
    0.0783775922876842, 0.0882057647363221, 0.126874479704313, 0.0444900553037874, 
    0.011210924284355, 0.0485436288069594), SE = c(0.0276312788788636, 
    0.00951096809292287, 0.0247673427299119, 0.0283463950228221, 
    0.0300166953905415, 0.0119050560847632, 0.0411216577835893, 0.0225049995889141, 
    0.0369242451412551, 0.00851003846804108, 0.00416915647513517, 
    0.00759714096622745, 0.0277959944131186, 0.0170947208638228, 
    0.0137755404255046, 0.0253680054199854, 0.00408412159854824, 
    0.024487788430806, 0.0428177553751394, 0.0411713999641989, 0.0396727915867888, 
    0.0160614455837079, 0.030112140497714, 0.0178581517236676, 0.0419537076967664, 
    0.0337028333157947, 0.0387119000488496, 0.0289249324044521, 0.0217931177697776, 
    0.0347896821174935, 0.0121005224896259, 0.0373634519234081, 0.0323364444614249, 
    0.0192577634336539, 0.0131406572658402, 0.0217853857228293, 0.0178665408351603, 
    0.00434478754261805, 0.0201829837822752, 0.0163428577310708, 
    0.0183921733504504, 0.026455157794341, 0.00927681781303608, 0.00233763930819051, 
    0.0101220463168873), CI = c(0.0573037650986205, 0.0197245405776072, 
    0.0513643250511128, 0.0587868252099994, 0.0622508161579229, 0.0246895751860773, 
    0.0852810985849519, 0.0466725125406402, 0.0765761975511351, 0.0176487395856976, 
    0.00864630132966052, 0.0157555060429295, 0.0576453642089773, 
    0.0354522812028192, 0.0285687222842772, 0.052610023226951, 0.00846994978927434, 
    0.0507845649191804, 0.0887985897006044, 0.0853842575536569, 0.0822763340004107, 
    0.0333093994276579, 0.0624487571951513, 0.0370355399022581, 0.0870066644910416, 
    0.0698953983256852, 0.0802835669187016, 0.0599866383045376, 0.0451961600084025, 
    0.072149384784412, 0.025094947698802, 0.0774870566678828, 0.0670616812803235, 
    0.0399381569331217, 0.0272520551981834, 0.0451801247244766, 0.0370529378546465, 
    0.00901053787040049, 0.0418569464960043, 0.0338930125010548, 
    0.0381430329717212, 0.054864639256835, 0.0192389426183078, 0.0048479672037301, 
    0.0209918392486697)), .Names = c("Visual", "Audio", "Report", 
    "Prob", "SD", "SE", "CI"), row.names = c(NA, -45L), class = "data.frame")

Here is my code to generate the graph:

#####################################################
######### ADD THIS LINE TO CREATE THE HIGHLIGHT SUBSET
HighlightDataShort<-AggBar[AggBar$Report==gsub("\\s|\\(40ms)","", AggBar$Visual),]
HighlightDataLong <-AggBar[AggBar$Report==gsub("\\s|\\(90ms)", "", AggBar$Visual),]
HighlightDataShort<-AggBar[AggBar$Report==gsub("\\s|\\(40ms)","", AggBar$Visual),]
HighlightDataSIM <-AggBar[AggBar$Report==gsub("\\s|\\Circle Presentation", "", AggBar$Visual),]
#####################################################


prob.bar = ggplot(AggBar, aes(x = Report, y = Prob, fill = Report)) + theme_bw() + facet_grid(Audio~Visual)
Exp2BGrid <- prob.bar + geom_bar(position=position_dodge(.9), stat="identity", colour="black", width=0.8) + theme(legend.position = "none") + labs(x="Report", y="Probability of Report") + scale_fill_grey() +

  ######### ADD THIS LINE TO CREATE THE HIGHLIGHT SUBSET
  geom_bar(data=HighlightDataShort, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.80)+
  geom_bar(data=HighlightDataShort, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.60)+
  geom_bar(data=HighlightDataShort, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.40)+
  geom_bar(data=HighlightDataShort, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.20)+
  geom_bar(data=HighlightDataShort, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.0) +
  ######################################################

######### ADD THIS LINE TO CREATE THE HIGHLIGHT SUBSET
geom_bar(data=HighlightDataLong, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.80)+
  geom_bar(data=HighlightDataLong, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.60)+
  geom_bar(data=HighlightDataLong, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.40)+
  geom_bar(data=HighlightDataLong, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.20)+
  geom_bar(data=HighlightDataLong, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.0) +
  ######################################################
######### ADD THIS LINE TO CREATE THE HIGHLIGHT SUBSET
geom_bar(data=HighlightDataSIM, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.80)+
  geom_bar(data=HighlightDataSIM, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.60)+
  geom_bar(data=HighlightDataSIM, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.40)+
  geom_bar(data=HighlightDataSIM, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.20)+
  geom_bar(data=HighlightDataSIM, position=position_dodge(.9), stat="identity", colour="black", size=.5, width=0.0) +
  ######################################################



labs(title = expression("Visual Condition")) +
  scale_fill_grey(start=.4) + # CODE TO LIGHTEN RELATIVE COLOUR SCHEME OF BARS
  # Move the y-axis label Source: http://stackoverflow.com/questions/8454243/move-axis-labels-ggplot
  theme(axis.title.y = element_text(vjust=2)) +
  theme(axis.title.x = element_text(vjust=-0.05)) +
  theme(text=element_text(size=30))+
  theme(axis.text.y =element_text(size=20))+ # THIS CHANGES SIZE OF VALUES ON Y-AXIS
  theme(axis.title.x=element_text(size=40))+
  theme(axis.title.y=element_text(size=40))+
  #theme(axis.text.x=element_text(size=30))+
  geom_errorbar(aes(ymin=Prob-CI, ymax=Prob+CI),
                width=.1, # Width of the error bars
                position=position_dodge(.9))+
  theme(plot.title = element_text(size = 40, vjust=1))+ #vjust respositions text
  scale_y_continuous(limits = c(0, 1), breaks = (seq(0,1,by = .25)))

Exp2BGrid

The generated graph is as it should be (note this is jpeg but when I manually save a PDF it is also as it should be):

However, when I save using:

ggsave(Exp2BGrid,file="Experiment_2A.pdf", width=22, height=13, path='/home/filelocation')

the following code is ignored:

theme(axis.title.y = element_text(vjust=2)) +
  theme(axis.title.x = element_text(vjust=-0.05))

The output when saved via ggsave ignores the repositioning of the axis text and it overlaps values etc.

Any insight would be most welcome.

score 0 · Accepted Answer · answered Nov 12 '15 at 18:05

0

The problem was actually down to redundant coding. ggsave doesn't appear to play well with repetition of axis.title code. I tidied it up as so:

theme(axis.title.y = element_text(size=40, vjust=2)) +
theme(axis.title.x = element_text(size=40, vjust=-0.05))

getting rid of the separate lines for size and ggsave works fine again.

answered Nov 12 '15 at 18:05

Docconcoct

2,040
4
28
52

I haven't studied it in detail, but your whole code doesn't follow the principles behind ggplot2 ("the grammar of graphics") and is overly complicated as a result. How many calls to `geom_bar` are that? There should be only one. – Roland Nov 12 '15 at 18:08
@Roland definitely non-standard. Notice the stripes on the bars in OP's image - they are created by overplotting bars of various widths. The data, at least, is surprisingly tidy given the length of the code. – Gregor Thomas Nov 12 '15 at 18:09
@Roland @ Gregor Yeah, I'm using a hack I came up with some time ago to allow for patterns on desired bars. Here's the hack with different pattern options if you're interested: http://stackoverflow.com/a/20426482/2653210 – Docconcoct Nov 12 '15 at 18:11
If you insist on using bar plots (blech!) and on hatching them (yuck!) at least write a function that achieves this. I don't know how often you've used this hack (at least twice apparently), but the time for writing a function would be well invested if you need that many lines otherwise. – Roland Nov 12 '15 at 18:50
@Roland you certainly say it how you see it don't you? We clearly have differing views on what is acceptable in data presentation. Yes, I do plan on writing a function for those hacks but it's not at the top of my list of priorities at the moment. Thanks for your uninvited yet constructive feedback on my hack. All the best. – Docconcoct Nov 12 '15 at 19:00

Problems exporting ggplot2 graph as PDF using ggsave

1 Answers1