R geom_text with dodged barplot with xaxis date

Question

I try to add text as in Position geom_text on dodged barplot But it doesn't work with a xaxis Date within ggplot:

Data<-as.data.frame(             
  cbind( 
    as.POSIXct(c("2017-06-26", "2017-06-26" ,"2017-06-26" ,"2017-07-14", "2017-07-14",
    "2017-07-14", "2017-07-14" ,"2017-07-14", "2017-09-14", "2017-09-14"),origin = "1970-01-01", tz = "GMT"),
  c("CHLOROPHYTA","CYANOBACTERIA","INDETERMINES","BACILLARIOPHYTA","CHLOROPHYTA", "CYANOBACTERIA" ,   
    "HAPTOPHYTA","HETEROKONTOPHYTA","CHLOROPHYTA","CYANOBACTERIA" ),
  c(29,8637,3233,97,2816, 63721,282, 18,2001,76593)))
  colnames(Data)<-c("Date","Embranchement","NbCel")

  Data$Date<-as.POSIXct(as.numeric(as.character(Data$Date)), origin = "1970-01-01", tz = "GMT")

  my_plot<-
    ggplot(data=Data, aes(x=Date, y=NbCel, fill=Embranchement,width=1150000)) +
    geom_bar(position = "dodge", stat="identity")+
    geom_text(aes(label=NbCel), vjust=1.6, color="black",
              position = position_dodge(0.9), size=3.5)+
    scale_x_discrete(labels = Data$Date) +
    scale_fill_brewer(palette="Paired")#+
  #axis(1,at=Data$Date,labels = Data$Date)
  print(my_plot)

What I get is :

How do I add text in the middle of each bar?

One additional question is how do I add each x thick mark for each date ?

Thanks!

Answer 1

I am not completely sure what you want to achieve in detail, but here are my 2 cents. You could use the following code:

Your first problem is that NbCel is a factor and is not numeric. You can easily see that on the Y-axis on your graph.

# convert NbCel from factor to integer 
Data$NbCel <- as.integer(as.character(Data$NbCel))

Having fixed this, we work with the graph itself.

The X-axis is a mixture of "date" information an "embranchement". Maybe it is better to separate out those two in different panels (facets) using facet_wrap.

# use a facet-wrap (separate plots in one graph)

ggplot(data=Data, aes(x=Embranchement, y=NbCel, fill=Embranchement)) +
  geom_bar(position = position_dodge(), stat="identity")+
  facet_wrap(~Date ,  ncol = 3)   +         ### plot per date
  scale_fill_brewer(palette="Paired") +
  geom_text(aes(label=NbCel, group = Embranchement), vjust=-.5, color="black",
            position = position_dodge(width = 10), size=2.5) + 
  theme(axis.text.x = element_text(angle = 60, hjust = 1, size = 6))  # make X labels readible

Please let me know whether this is what you want.

ADDITION

Again, you want to combine different bars from "embranchement" and different "dates" on the X-axis. A more natural thing to do in this case might be a line graph, this way you can superimpose frequency information concerning different data points on 1 date.

Your dataset is not full, i.e. not every date contains all "embranchements".

# indicates measurement not available on every date, hence no lines
with(Data, table(Embranchement, Date))
Embranchement      2017-06-26 2017-07-14 2017-09-14
  BACILLARIOPHYTA           0          1          0
  CHLOROPHYTA               1          1          1
  CYANOBACTERIA             1          1          1
  HAPTOPHYTA                0          1          0
  HETEROKONTOPHYTA          0          1          0
  INDETERMINES              1          0          0

Assuming that absent readings are null-readings, I filled up the dataset, yielding Data2:

Data2 <- rbind(Data, 
               data.frame(Date = as.Date("2017-06-26", format="%Y-%m-%d"), Embranchement = "BACILLARIOPHYTA", NbCel = 0),
               data.frame(Date = as.Date("2017-06-26", format="%Y-%m-%d"), Embranchement = "HAPTOPHYTA", NbCel = 0),
               data.frame(Date = as.Date("2017-06-26", format="%Y-%m-%d"), Embranchement = "HETEROKONTOPHYTA", NbCel = 0),
               data.frame(Date = as.Date("2017-07-14", format="%Y-%m-%d"), Embranchement = "INDETERMINES", NbCel = 0),
               data.frame(Date = as.Date("2017-09-14", format="%Y-%m-%d"), Embranchement = "BACILLARIOPHYTA", NbCel = 0),
               data.frame(Date = as.Date("2017-09-14", format="%Y-%m-%d"), Embranchement = "HAPTOPHYTA", NbCel = 0),
               data.frame(Date = as.Date("2017-09-14", format="%Y-%m-%d"), Embranchement = "HETEROKONTOPHYTA", NbCel = 0),
               data.frame(Date = as.Date("2017-09-14", format="%Y-%m-%d"), Embranchement = "INDETERMINES", NbCel = 0)
               )

Now we may create a line graph:

ggplot(data=Data2, aes(x=Date, y=NbCel, col=Embranchement, linetype =Embranchement) ) +
  geom_line( size=1.1  ) +
  scale_fill_brewer(palette="Paired") +
  geom_text(aes(label=NbCel), vjust=-.75, color="black", size=2.5) 

Yielding the following graph, with "date" on X-axis and a line per "Embranchement".

Even if this isn't what you want, I hope it helps you.

ADDITION 2 (recap of facets but better)

(this answer needs editing, later ;-)

Try this code to get clean numbers on the bars in the facets. The most important change is in geom_text: position_dodge

ggplot(data=Data, aes(x=Embranchement, y=NbCel, fill=Embranchement)) +
  geom_bar(position = position_dodge(), stat="identity")+
  facet_wrap(~Date ,  ncol = 3)   +         ### plot per date
  scale_fill_brewer(palette="Paired") +
  theme_light() +
  geom_text(aes(label=NbCel, group = Embranchement), vjust=-.5, color="black",
            position = position_dodge(width = 1), size=2.5) +
  theme(axis.text.x = element_text(angle = 60, hjust = 1, size = 6)) 

You will get this graph:

I think this is closer to what you want. Let me know.

Answer 2

Your first problem is that NbCel is a factor and is not numeric. You can easily see that on the Y-axis on your graph.

# convert NbCel from factor to integer 
Data$NbCel <- as.integer(as.character(Data$NbCel))

Now after using some of the code from @KoenV, we now can

ggplot(data=Data, aes(x=as.character(Date), y=NbCel)) +
       geom_bar(aes(fill=Embranchement), position = position_dodge(), stat="identity") +
       scale_fill_brewer(palette="Paired") +
       geom_text(aes(label=NbCel, y = NbCel + 1200, group = Embranchement), color="black", 
                 position = position_dodge(0.9), size=3.5)