using ggplot to plot curvilinear fit - r

How do I plot a curve for a line of best fit using ggplot? My best guess is that I need to change the stat_smooth parameter somehow but I have no clue how. My goal is something like the black line in the image below.
vv<-structure(list(X = 16:19, school = structure(c(3L, 3L, 3L, 3L), .Label = c("UCB", "UCD", "UIUC"), class = "factor"), year = 2009:2012, mean = c(15.60965, 16.785, 16.77725, 15.91729), sd = c(6.483547,6.852999, 6.327013, 6.74991)), .Names = c("X", "school", "year", "mean", "sd"), row.names = 16:19, class = "data.frame")
ggplot(vv, aes(x = year, y = mean)) +
ggtitle("scores")+
geom_point() +
stat_smooth(method = "lm", col = "red")

You can try changing the formula:
ggplot(vv, aes(x = year, y = mean)) +
ggtitle("scores")+
geom_point() +
stat_smooth(method = "lm", formula = y ~ splines::bs(x, 3), col = "red")

Or simply this (with loess) although you will get some warnings due to span too small for your very small data, but it works:
ggplot(vv, aes(x = year, y = mean)) +
ggtitle("scores") +
geom_point(size=3) +
stat_smooth(col = "red")

Related

How do I create a frequency stacked bar chart however have percentage labels on the bars and frequencies on the y axis, in R?

I started with the code below, however it is not showing the right output. I would just like a normal frequency stacked bar chart to show percentages on the bars but frequencies on the y axis. Could anyone offer any suggestions please?
ggplot(data = df, mapping = aes(x = Family_Size, y = Freq, fill = Survived)) + geom_bar(stat = "identity") +
geom_text(aes(label = paste0(df$Percentage),y=Percentage),size = 3) +
theme(plot.title = element_text(hjust = 0.5))
<table><tbody><tr><th>Survived</th><th>Family_Size</th><th>Frequency</th><th>Percentage</th></tr><tr><td>Yes</td><td>1</td><td>20</td><td>20%</td></tr><tr><td>No</td><td>1</td><td>80</td><td>80%</td></tr><tr><td>Yes</td><td>2</td><td>40</td><td>40%</td></tr><tr><td>No</td><td>2</td><td>60</td><td>60%</td></tr></tbody></table>
Are you looking for something like that ?
ggplot(df, aes(x = Family_Size, y = Frequency, fill = Survived))+
geom_col()+
scale_y_continuous(breaks = seq(0,100, by = 20))+
geom_text(aes(label = Percentage), position = position_stack(0.5))
EDIT: Formatting percentages with two decimales
ggplot(df, aes(x = Family_Size, y = Frequency, fill = Survived))+
geom_col()+
scale_y_continuous(breaks = seq(0,100, by = 20))+
geom_text(aes(label = paste(format(round(Frequency,2),nsmall = 2),"%")), position = position_stack(0.5))
Reproducible example
structure(list(Survived = c("Yes", "No", "Yes", "No"), Family_Size = c(1L,
1L, 2L, 2L), Frequency = c(20L, 80L, 40L, 60L), Percentage = c("20%",
"80%", "40%", "60%")), row.names = c(NA, -4L), class = c("data.table",
"data.frame"))

Positioning labels and color coding in sunburst - R

This is what is the output.I have a data set which contains unit, weight of each unit and compliance score for each unit in year 2016.
I was not able to add the table but here is the screenshot for the data in csv
I have named the columns in the data as unit, weight and year(which is compliance score) .
I want to create a sunburst chart where the first ring will be the unit divided based on weight and the second ring will be the same but will have labels compliance score.
The colour for each ring will be different.
I was able to do some code with the help from an online blog and the output I have gotten is similar to what I want but I am facing difficulty in positioning of the labels and also the colour coding for each ring
#using ggplot
library(ggplot2) # Visualisation
library(dplyr) # data wrangling
library(scales) # formatting
#read file
weight.eg = read.csv("Dummy Data.csv", header = FALSE, sep =
";",encoding = "UTF-8")
#change column names
colnames(weight.eg) <- c ("unit","weight","year")
#as weight column is factor change into integer
weight.eg$weight = as.numeric(levels(weight.eg$weight))
[as.integer(weight.eg$weight)]
weight.eg$year = as.numeric(levels(weight.eg$year))
[as.integer(weight.eg$year)]
#Nas are introduced, remove
weight.eg <- na.omit(weight.eg)
#Sum of the total weight
sum_total_weight = sum(weight.eg$weight)
#First layer
firstLevel = weight.eg %>% summarize(total_weight=sum(weight))
sunburst_0 = ggplot(firstLevel) # Just a foundation
#this will generate a bar chart
sunburst_1 =
sunburst_0 +
geom_bar(data=firstLevel, aes(x=1, y=total_weight),
fill='darkgrey', stat='identity') +
geom_text(aes(x=1, y=sum_total_weight/2, label=paste("Total
Weight", comma(total_weight))), color='black')
#View
sunburst_1
#this argument is used to rotate the plot around the y-axis which
the total weight
sunburst_1 + coord_polar(theta = "y")
sunburst_2=
sunburst_1 +
geom_bar(data=weight.eg,
aes(x=2, y=weight.eg$weight, fill=weight.eg$weight),
color='white', position='stack', stat='identity', size=0.6)
+
geom_text(data=weight.eg, aes(label=paste(weight.eg$unit,
weight.eg$weight), x=2, y=weight.eg$weight), position='stack')
sunburst_2 + coord_polar(theta = "y")
sunburst_3 =
sunburst_2 +
geom_bar(data=weight.eg,
aes(x=3, y=weight.eg$weight,fill=weight.eg$weight),
color='white', position='stack', stat='identity',
size=0.6)+
geom_text(data = weight.eg,
aes(label=paste(weight.eg$year),x=3,y=weight.eg$weight),position =
'stack')
sunburst_3 + coord_polar(theta = "y")
sunburst_3 + scale_y_continuous(labels=comma) +
scale_fill_continuous(low='white', high='darkred') +
coord_polar('y') + theme_minimal()
Output for dput(weight.eg)
structure(list(unit = structure(2:7, .Label = c("", "A", "B",
"C", "D", "E", "F", "Unit"), class = "factor"), weight = c(30,
25, 10, 17, 5, 13), year = c(70, 80, 50, 30, 60, 40)), .Names =
c("unit",
"weight", "year"), row.names = 2:7, class = "data.frame", na.action
= structure(c(1L,
8L), .Names = c("1", "8"), class = "omit"))
output for dput(firstLevel)
structure(list(total_weight = 100), .Names = "total_weight", row.names
= c(NA,
-1L), na.action = structure(c(1L, 8L), .Names = c("1", "8"), class =
"omit"), class = "data.frame")
So I think I might have some sort of solution for you. I wasn't sure what you wanted to color-code on the outer ring; from your code it seems you wanted it to be the weight again, but it was not obvious to me. For different colour scales per ring, you could use the ggnewscale package:
library(ggnewscale)
For the centering of the labels you could write a function:
cs_fun <- function(x){(cumsum(x) + c(0, cumsum(head(x , -1))))/ 2}
Now the plotting code could look something like this:
ggplot(weight.eg) +
# Note: geom_col is equivalent to geom_bar(stat = "identity")
geom_col(data = firstLevel,
aes(x = 1, y = total_weight)) +
geom_text(data = firstLevel,
aes(x = 1, y = total_weight / 2,
label = paste("Total Weight:", total_weight)),
colour = "black") +
geom_col(aes(x = 2,
y = weight, fill = weight),
colour = "white", size = 0.6) +
scale_fill_gradient(name = "Weight",
low = "white", high = "darkred") +
# Open up new fill scale for next ring
new_scale_fill() +
geom_text(aes(x = 2, y = cs_fun(weight),
label = paste(unit, weight))) +
geom_col(aes(x = 3, y = weight, fill = weight),
size = 0.6, colour = "white") +
scale_fill_gradient(name = "Another Weight?",
low = "forestgreen", high = "white") +
geom_text(aes(label = paste0(year), x = 3,
y = cs_fun(weight))) +
coord_polar(theta = "y")
Which looks like this:

Using ggplot Geom_Ribbon in R to fill under a continuous line

I am trying to use geom_ribbon to fill an area under a geom_smooth line in ggplot and there are gaps under the curve where the color is not shaded. My data consists of six discrete values for proportion values on the y axis. Is there a way to use ymax in geom_ribbon differently to have the color meet the curved line better?
Here is the reproducible code for the data:
q1 <- structure(list(Session = 1:6, Counts = c(244L, 358L, 322L, 210L,
156L, 100L), Density_1000 = c(NA, NA, NA, NA, NA, NA), Proportion_Activity = c(0.175539568,
0.257553957, 0.231654676, 0.151079137, 0.112230216, 0.071942446
), Lifestage = structure(c(3L, 3L, 3L, 3L, 3L, 3L), .Label = c("Adult",
"Nymph", "Larvae"), class = "factor")), .Names = c("Session",
"Counts", "Density_1000", "Proportion_Activity", "Lifestage"), row.names = 13:18, class = "data.frame")
Here is the ggplot code:
ggplot(q1,aes(x=Session, y=Proportion_Activity, col = Lifestage,fill=Lifestage))
+ geom_smooth(method = 'loess')
+ geom_ribbon(data = q1,aes(x = Session, ymin=0, ymax=Proportion_Activity, alpha=0.5))
You can just use the area geom with the stat_smooth layer. For example
ggplot(q1,aes(x=Session, y=Proportion_Activity, col = Lifestage,fill=Lifestage)) +
geom_smooth(method = 'loess') +
stat_smooth(se=FALSE, geom="area", method = 'loess', alpha=.5)
Thou I really think smoothing should be used when you have a lot of data and want to show a general pattern. Using it like this to "smooth" the line to make it look pretty doesn't make it clear that you have modeled the results and shows data in places where you did not observe it.
You can do something like this.
p1 <- ggplot(q1,aes(x=Session, y=Proportion_Activity)) +
geom_smooth(method = 'loess', aes(color = Lifestage))
g1 <- ggplot_build(p1)
p2 <- data.frame(Session = g1$data[[1]]$x,
Proportion_Activity = g1$data[[1]]$y,
Lifestage = structure(g1$data[[1]]$group, .Label = c("Larvae", "Nymph", "Adult"), class = "factor"))
p1 + geom_ribbon(data = p2, aes(x = Session, ymin = 0, ymax = Proportion_Activity, fill = Lifestage), alpha = 0.5)
You can also use geom_line instead of geom_smooth.
geom_line(stat = "smooth", method = 'loess', alpha = 0.5, aes(color = Lifestage))
And remove the color from geom_smooth/geom_line if you want. Just add guides(color = FALSE) or fill if you want to remove that.

Plot confusion matrix in R using ggplot

I have two confusion matrices with calculated values as true positive (tp), false positives (fp), true negatives(tn) and false negatives (fn), corresponding to two different methods. I want to represent them as
I believe facet grid or facet wrap can do this, but I find difficult to start.
Here is the data of two confusion matrices corresponding to method1 and method2
dframe<-structure(list(label = structure(c(4L, 2L, 1L, 3L, 4L, 2L, 1L,
3L), .Label = c("fn", "fp", "tn", "tp"), class = "factor"), value = c(9,
0, 3, 1716, 6, 3, 6, 1713), method = structure(c(1L, 1L, 1L,
1L, 2L, 2L, 2L, 2L), .Label = c("method1", "method2"), class = "factor")), .Names = c("label",
"value", "method"), row.names = c(NA, -8L), class = "data.frame")
This could be a good start
library(ggplot2)
ggplot(data = dframe, mapping = aes(x = label, y = method)) +
geom_tile(aes(fill = value), colour = "white") +
geom_text(aes(label = sprintf("%1.0f",value)), vjust = 1) +
scale_fill_gradient(low = "white", high = "steelblue")
Edited
TClass <- factor(c(0, 0, 1, 1))
PClass <- factor(c(0, 1, 0, 1))
Y <- c(2816, 248, 34, 235)
df <- data.frame(TClass, PClass, Y)
library(ggplot2)
ggplot(data = df, mapping = aes(x = TClass, y = PClass)) +
geom_tile(aes(fill = Y), colour = "white") +
geom_text(aes(label = sprintf("%1.0f", Y)), vjust = 1) +
scale_fill_gradient(low = "blue", high = "red") +
theme_bw() + theme(legend.position = "none")
It is a very old question, still it seems there is a quite straight forward solution to that using ggplot2 which hasn't been mentioned.
Hope it might be helpful to someone:
cm <- confusionMatrix(factor(y.pred), factor(y.test), dnn = c("Prediction", "Reference"))
plt <- as.data.frame(cm$table)
plt$Prediction <- factor(plt$Prediction, levels=rev(levels(plt$Prediction)))
ggplot(plt, aes(Prediction,Reference, fill= Freq)) +
geom_tile() + geom_text(aes(label=Freq)) +
scale_fill_gradient(low="white", high="#009194") +
labs(x = "Reference",y = "Prediction") +
scale_x_discrete(labels=c("Class_1","Class_2","Class_3","Class_4")) +
scale_y_discrete(labels=c("Class_4","Class_3","Class_2","Class_1"))
A slightly more modular solution based on MYaseen208's answer. Might be more effective for large datasets / multinomial classification:
confusion_matrix <- as.data.frame(table(predicted_class, actual_class))
ggplot(data = confusion_matrix
mapping = aes(x = Var1,
y = Var2)) +
geom_tile(aes(fill = Freq)) +
geom_text(aes(label = sprintf("%1.0f", Freq)), vjust = 1) +
scale_fill_gradient(low = "blue",
high = "red",
trans = "log") # if your results aren't quite as clear as the above example
Here's another ggplot2 based option; first the data (from caret):
library(caret)
# data/code from "2 class example" example courtesy of ?caret::confusionMatrix
lvs <- c("normal", "abnormal")
truth <- factor(rep(lvs, times = c(86, 258)),
levels = rev(lvs))
pred <- factor(
c(
rep(lvs, times = c(54, 32)),
rep(lvs, times = c(27, 231))),
levels = rev(lvs))
confusionMatrix(pred, truth)
And to construct the plots (substitute your own matrix below as needed when setting up "table"):
library(ggplot2)
library(dplyr)
table <- data.frame(confusionMatrix(pred, truth)$table)
plotTable <- table %>%
mutate(goodbad = ifelse(table$Prediction == table$Reference, "good", "bad")) %>%
group_by(Reference) %>%
mutate(prop = Freq/sum(Freq))
# fill alpha relative to sensitivity/specificity by proportional outcomes within reference groups (see dplyr code above as well as original confusion matrix for comparison)
ggplot(data = plotTable, mapping = aes(x = Reference, y = Prediction, fill = goodbad, alpha = prop)) +
geom_tile() +
geom_text(aes(label = Freq), vjust = .5, fontface = "bold", alpha = 1) +
scale_fill_manual(values = c(good = "green", bad = "red")) +
theme_bw() +
xlim(rev(levels(table$Reference)))
# note: for simple alpha shading by frequency across the table at large, simply use "alpha = Freq" in place of "alpha = prop" when setting up the ggplot call above, e.g.,
ggplot(data = plotTable, mapping = aes(x = Reference, y = Prediction, fill = goodbad, alpha = Freq)) +
geom_tile() +
geom_text(aes(label = Freq), vjust = .5, fontface = "bold", alpha = 1) +
scale_fill_manual(values = c(good = "green", bad = "red")) +
theme_bw() +
xlim(rev(levels(table$Reference)))
Here it is a reprex using cvms package i.e., Wrapper function for ggplot2 to make confusion matrix.
library(cvms)
library(broom)
library(tibble)
library(ggimage)
#> Loading required package: ggplot2
library(rsvg)
set.seed(1)
d_multi <- tibble("target" = floor(runif(100) * 3),
"prediction" = floor(runif(100) * 3))
conf_mat <- confusion_matrix(targets = d_multi$target,
predictions = d_multi$prediction)
# plot_confusion_matrix(conf_mat$`Confusion Matrix`[[1]], add_sums = TRUE)
plot_confusion_matrix(
conf_mat$`Confusion Matrix`[[1]],
add_sums = TRUE,
sums_settings = sum_tile_settings(
palette = "Oranges",
label = "Total",
tc_tile_border_color = "black"
)
)
Created on 2021-01-19 by the reprex package (v0.3.0)
Old question, but I wrote this function which I think makes a prettier answer. Results in a divergent color palette (or whatever you want, but default is divergent):
prettyConfused<-function(Actual,Predict,colors=c("white","red4","dodgerblue3"),text.scl=5){
actual = as.data.frame(table(Actual))
names(actual) = c("Actual","ActualFreq")
#build confusion matrix
confusion = as.data.frame(table(Actual, Predict))
names(confusion) = c("Actual","Predicted","Freq")
#calculate percentage of test cases based on actual frequency
confusion = merge(confusion, actual, by=c('Actual','Actual'))
confusion$Percent = confusion$Freq/confusion$ActualFreq*100
confusion$ColorScale<-confusion$Percent*-1
confusion[which(confusion$Actual==confusion$Predicted),]$ColorScale<-confusion[which(confusion$Actual==confusion$Predicted),]$ColorScale*-1
confusion$Label<-paste(round(confusion$Percent,0),"%, n=",confusion$Freq,sep="")
tile <- ggplot() +
geom_tile(aes(x=Actual, y=Predicted,fill=ColorScale),data=confusion, color="black",size=0.1) +
labs(x="Actual",y="Predicted")
tile = tile +
geom_text(aes(x=Actual,y=Predicted, label=Label),data=confusion, size=text.scl, colour="black") +
scale_fill_gradient2(low=colors[2],high=colors[3],mid=colors[1],midpoint = 0,guide='none')
}

Stacked Bar Graph Labels with ggplot2

I am trying to graph the following data:
to_graph <- structure(list(Teacher = c("BS", "BS", "FA"
), Level = structure(c(2L, 1L, 1L), .Label = c("BE", "AE", "ME",
"EE"), class = "factor"), Count = c(2L, 25L, 28L)), .Names = c("Teacher",
"Level", "Count"), row.names = c(NA, 3L), class = "data.frame")
and want to add labels in the middle of each piece of the bars that are the percentage for that piece. Based on this post, I came up with:
ggplot(data=to_graph, aes(x=Teacher, y=Count, fill=Level), ordered=TRUE) +
geom_bar(aes(fill = Level), position = 'fill') +
opts(axis.text.x=theme_text(angle=45)) +
scale_y_continuous("",formatter="percent") +
opts(title = "Score Distribution") +
scale_fill_manual(values = c("#FF0000", "#FFFF00","#00CC00", "#0000FF")) +
geom_text(aes(label = Count), size = 3, hjust = 0.5, vjust = 3, position = "stack")
But it
Doesn't have any effect on the graph
Probably doesn't display the percentage if it did (although I'm not entirely sure of this point)
Any help is greatly appreciated. Thanks!
The y-coordinate of the text is the actual count (2, 25 or 28), whereas the y-coordinates in the plot panel range from 0 to 1, so the text is being printed off the top.
Calculate the fraction of counts using ddply (or tapply or whatever).
graph_avgs <- ddply(
to_graph,
.(Teacher),
summarise,
Count.Fraction = Count / sum(Count)
)
to_graph <- cbind(to_graph, graph_avgs$Count.Fraction)
A simplified version of your plot. I haven't bothered to play about with factor orders so the numbers match up to the bars yet.
ggplot(to_graph, aes(Teacher), ordered = TRUE) +
geom_bar(aes(y = Count, fill = Level), position = 'fill') +
scale_fill_manual(values = c("#FF0000", "#FFFF00","#00CC00", "#0000FF")) +
geom_text(
aes(y = graph_avgs$Count.Fraction, label = graph_avgs$Count.Fraction),
size = 3
)

Resources