How to Get Discrete Factor Levels to Be Treated as Continuous

How do I get discrete factor levels to be treated as continuous?

I think you can do this simply by transforming the variable to numeric:

mdf$variable <- as.numeric(as.character(mdf$variable))

g <- ggplot(mdf,aes(variable,value,group=variable,colour=t))
g +
    geom_point() +
    #scale_x_continuous() +
    opts()

R: Factor level in ggplot treated as continuous data set

Does this resolve your issue? I added a variable that separates each change in CiPA, even if it's going back to one already used. That way it doesn't connect non-contiguous sections which share a CiPA level.

EDIT - here showing full code that works verbatim on my computer.

EDIT #2 - added adjusted line for text to ignore the CiPA_grp grouping variable which newdf won't have. aes(group = 1) in that layer will specify that it should put all elements (just one in actuality) in the same group, rather than looking to the CiPA_grp variable for that.

library(dplyr)
cidf %>%   # using the "bad" version
  arrange(drug, dose) %>%
  group_by(drug) %>%
  mutate(CiPA_grp = cumsum(as.numeric(CiPA) != lag(as.numeric(CiPA), default = Inf))) %>%
  ungroup() %>%

ggplot(aes(dose, value, group=interaction(drug, CiPA_grp))) + 
  scale_color_manual(values = c("2" = "#e82929", "1"="#337cb4", "0"="#44ae52")) + 
  scale_fill_manual(values = c("2" = "#e82929", "1"="#337cb4", "0"="#44ae52"), name="fill") + 
  geom_line(aes(color=CiPA)) + 
  geom_ribbon(aes(ymin=lower, ymax=upper, fill = CiPA), alpha=0.3) + 
  geom_text(data=newdf, aes(label=drug, color=CiPA, group = 1), hjust=-0.2, vjust=0.5, size=3, show.legend=F)  +
  coord_cartesian(xlim=c(0,max(cidf$dose)*1.2)) + 
  xlab(~"Concentration (\u00D7"~C[max]*")") + 
  ylab(~"qNet ("*mu*"C/"*mu*"F)") + 
  theme_bw() + 
  theme(legend.position="none")

Plotting with ggplot2: Error: Discrete value supplied to continuous scale on categorical y-axis

As mentioned in the comments, there cannot be a continuous scale on variable of the factor type. You could change the factor to numeric as follows, just after you define the meltDF variable.

meltDF$variable=as.numeric(levels(meltDF$variable))[meltDF$variable]

Then, execute the ggplot command

  ggplot(meltDF[meltDF$value == 1,]) + geom_point(aes(x = MW, y =   variable)) +
     scale_x_continuous(limits=c(0, 1200), breaks=c(0, 400, 800, 1200)) +
     scale_y_continuous(limits=c(0, 1200), breaks=c(0, 400, 800, 1200))

And you will have your chart.

Hope this helps

ggplot 2 Error: Discrete value supplied to continuous scale

moroccostats has the structure

> str(moroccostats)
'data.frame':   91 obs. of  38 variables:
 $ year            : Factor w/ 91 levels "1960","1961",..: 1 2 3 4 5 6 7 8 9 10 ...
 $ periodframe     : Factor w/ 4 levels "0","Phase 1 (1965 to 1985)",..: 1 1 1 1 1 2 2 2 2 2 ...
 $ pop_t           : num  11635 11897 12177 12473 12785 ...
 $ pop_ur          : num  3395 3547 3703 3862 4026 ...
 ...

When you melt that by year to get moroccostats.f, you get

> str(moroccostats.f)
'data.frame':   3367 obs. of  3 variables:
 $ year    : Factor w/ 91 levels "1960","1961",..: 1 2 3 4 5 6 7 8 9 10 ...
 $ variable: Factor w/ 37 levels "periodframe",..: 1 1 1 1 1 1 1 1 1 1 ...
 $ value   : chr  "0" "0" "0" "0" ...

Note that value is a character because periodframe was a factor. I'm guessing what changed was that periodframe went from being a number to a character (the "Phase 1 (1965 to 1985)" bits).

Since you are only plotting the pop_t and pop_ur and pop_ru columns, pull those out before melting

morstats.pop <- melt(moroccostats[c("year","pop_t","pop_ur","pop_ru")], id="year")

then you don't have to worry about coercions to other types dues to irrelevant columns.

How to mutate numeric to factor and keep levels 0 and 1 and not 1 and 2

With ggplot, stat_smooth, we need a continuous axis, which means not a factor. From the examples in the ?geom_smooth documentation:

# To fit a logistic regression, you need to coerce the values to

# a numeric vector lying between 0 and 1.

ggplot(rpart::kyphosis, aes(Age, as.numeric(Kyphosis) - 1)) +

...

Following these instructions (this process would be the same no matter what your factor levels are, as long as there are 2 levels):

ggplot(data, aes(x = age, y = as.numeric(factor1) - 1)) + 
  geom_point() + 
  stat_smooth(method="glm", se=FALSE, method.args = list(family=binomial))

Your problem is not with the factor, but with the subsequent methods. If you have more problems, I'd encourage you to ask more questions illustrating those problems. But probably this similar technique will work--binary data can be left as integer 1s and 0s, there's often no reason to convert such columns to factors at all, and many methods that expect binary data may not expect factors.

StatBin requires a continuous x variable the x variable is discrete

A histogram works with a continuous x axis, as the error states. Your x-axis is on pclass, which is a factor (which R treats as discrete).

If you want to get a count by pclass type, you want a geom_bar instead:

ggplot(train, aes(x = pclass, fill = factor(survived))) +
    geom_bar(stat = "count") +
    xlab("pclass") +
    ylab("total count") +
    labs(fill = "survived")

---

Related Topics