03-08-2017 03:34 PM
Here's is an example of creating a histogram with a density distribution line.
hist(dataset$Age, main = "Customer Count by Age", ylab="Customer Count", xlab="Customer Age", xlim = c(18, 100), border="black", breaks=20, col=c("lightyellow", "lightblue"), las=1, probability = TRUE ) lines(density(dataset$Age),lty="dotdash", lwd=4, col="red")
Here's an example of creating a barblot
barplot(dataset$'Sales Revenue', names.arg = dataset$'Age Group', main = "Sales Revenue by Customer Age Group", col = c("red","yellow","orange","blue", "green") ) minValue <- 0 maxValue <- max(as.vector(dataset$'Sales Revenue')) yTicks <- seq(from=minValue, to = maxValue, length.out = 10) yTicks <- pretty(yTicks) yTickLabels <- paste("$",format(yTicks/1000, , big.mark=","), "K",sep="") axis(2, at=yTicks, labels = yTickLabels, lty = 1, las=1, cex.axis=0.7 )
Of course, these are simple examples using the built-in R graphics functionality. You can also use a richer graphics package such as lattice or ggplot2 to create some really detailed charts and graphs.
Is this what you are looking for?
03-09-2017 08:45 AM
Thanks Ted, this is along the same lines, but I am interested more in an "xbar" chart and also looking for a simple R script to create a Pareto chart for my customers.
Thank you for the scripts for the Histogram and Bar Plot, I will also be adding this to my arsenal.
03-14-2017 01:33 PM
RE: also looking for a simple R script to create a Pareto chart
In order to use any chart -that is part of a R Visualization package library - it's logical to first make sure it's installed.
Now considering the huge amount of R package visualization libraries, there may be a few ways to reproduce what you want. Note that I don't consider myself an R guru at this stage, still here's a couple of basic samples.
library(qcc) defect <- c(80, 27, 66, 94, 33) names(defect) <- c("price code", "schedule date", "supplier code", "contact num.", "part num.") pareto.chart(defect, ylab = "Error frequency", col=heat.colors(length(defect)))
library(ggplot2) counts <- c(80, 27, 66, 94, 33) defects <- c("price code", "schedule date", "supplier code", "contact num.", "part num.") dat <- data.frame( count = counts, defect = defects, stringsAsFactors=FALSE ) dat <- dat[order(dat$count, decreasing=TRUE), ] dat$defect <- factor(dat$defect, levels=dat$defect) dat$**bleep** <- cumsum(dat$count) dat ggplot(dat, aes(x=defect)) + geom_bar(aes(y=count), fill="blue", stat="identity") + geom_point(aes(y=**bleep**)) + geom_path(aes(y=**bleep**, group=1))
I haven't played much with x-Bar chart though I'll check if I can find something... Anyway hope this helps
03-16-2017 02:09 PM
RE : xBarChart
In true life scenario, I'd say the following data would probably be obtained as a result
of fancy mathemathic algorithm formula... though I'll leave that to Math PhDs...
So for for simplicity I just added a list of subgroup manually...
So here we go...
#declare qcc library: library(qcc) # Load a mock list of 10 subgroup data manually: sg1 <- c(1.397742,1.399917,1.278918,1.279828) # Fill in subgroup 1 data! sg2 <- c(1.283877,1.307215,1.341566,1.396107) # Fill in subgroup 2 data! sg3 <- c(1.313634,1.278839,1.242498,1.331201) # Fill in subgroup 3 data! sg4 <- c(1.245943,1.303432,1.390168,1.298949) # Fill in subgroup 4 data! sg5 <- c(1.188624,1.226905,1.217450,1.284127) # Fill in subgroup 5 data! sg6 <- c(1.287875,1.283185,1.234186,1.314055) # Fill in subgroup 6 data! sg7 <- c(1.276711,1.284376,1.305309,1.249184) # Fill in subgroup 7 data! sg8 <- c(1.312219,1.297509,1.272367,1.371223) # Fill in subgroup 8 data! sg9 <- c(1.378350,1.312981,1.381944,1.268875) # Fill in subgroup 9 data! sg10 <- c(1.332196,1.268824,1.299608,1.329053) # Fill in subgroup 10 data! # Include those subgroups into a my.data mock list through rbind my.data <- rbind(sg1,sg2,sg3,sg4,sg5,sg6,sg7,sg8,sg9,sg10) # Draw the R Chart and calculate relevant metrics q1 <- qcc(my.data, type="R", nsigmas=3)
which should generate similar R chart:
then add following to ldraw the X-BAR chart...
# Draw the X-Bar Chart and calculate relevant metrics q2 <- qcc(my.data, type="xbar", nsigmas=3)
which should generate following X-Bar Chart:
# Establish the LSL and USL as set by customer specs, then # draw the process capability chart and calculate metrics: lsl <- 1.31 # Fill in a mock LSL here! usl <- 1.32 # Fill in a mock USL here! process.capability(q2, spec.limits=c(lsl,usl))
which should end up with following:
Hope this helps...
03-17-2017 09:01 AM
Thank you for all of this information! I am going to try and take some time this weekend to put into play your examples.
Much appreciated! Ill reply back with any progress I can make.
03-20-2017 06:52 AM
I have been able to create the R and XBar charts in PowerBI R Visuals, but still need some adjustments as it doesnt look right to me (data not displaying as expected).
months<- c("Jan", "Feb", "Mar", "Apr","May", "Jun","Jul","Aug","Sep","Oct","Nov","Dec")
qcc(dataset, type="R", nsigmas=1, labels=months, xlab= "Month", ylab = "Service Level %", title = "Phone Call SVL", digits=3)
Any idea how to get the Y axis to show as percentages? My data is in percentages as you can see when selecting a simple Line Chart from the PowerBI Visuals.
Thanks for all your assistance!
Sorry for late response. For the sake of simplicity, I mainly used mock data which I entered
manually. The idea was basically to create a series of 4 weeks subsets over a 12 months period sort if thing...
We both agree that a real life scenario would have gone through a much fancier scenario such as querying data
from a server and even adding conversion at some stage. I used numeric for simplicity.
Also I noticed that even if qcc library offers a wide variety of statisitic chart [R, xBar,...], it doesn't seem to have
any qcc - y axis formatting options parameter as you would probably find using scaling_y_continuous in ggplot2 library.
So in this case I don't believe y axis percentage formatting could be done in one shot.
My best bet would be to define the qcc : q1 R Chart and q2 xBar in respectice class with a plot=False attribute
library(qcc) Jan <- c(0.837742,0.839917,0.728918,0.729828) # Fill in subgroup January data! Feb <- c(0.783877,0.807215,0.841566,0.836107) # Fill in subgroup February data! Mar <- c(0.813634,0.728839,0.742498,0.831201) # Fill in subgroup March data! Apr <- c(0.745943,0.803432,0.830168,0.798949) # Fill in subgroup April data! May <- c(0.688624,0.726905,0.717450,0.784127) # Fill in subgroup May data! Jun <- c(0.787875,0.783185,0.714186,0.814055) # Fill in subgroup June data! Jul <- c(0.726711,0.784376,0.805309,0.749184) # Fill in subgroup July data! Aug <- c(0.812219,0.797509,0.722367,0.871223) # Fill in subgroup August data! Sep <- c(0.878350,0.812981,0.881944,0.768875) # Fill in subgroup September data! Oct <- c(0.832196,0.768824,0.799608,0.729053) # Fill in subgroup October data! Nov <- c(0.813634,0.728839,0.742498,0.831201) # Fill in subgroup November data! Dec <- c(0.726711,0.784376,0.805309,0.749184) # Fill in subgroup December data! # Include those subgroups into a my.data mock list through rbind dataset <- rbind(Jan,Feb,Mar,Apr,May,Jun,Jul,Aug,Sep,Oct,Nov,
Dec) months<- c("Jan", "Feb", "Mar", "Apr","May", "Jun","Jul","Aug","Sep","Oct","Nov","Dec")
# assign q1 q1 <- qcc(dataset, type="R", nsigmas=1, labels=months, xlab= "Month", ylab = "Service Level %", title = "Phone Call SVL", digits=3, label.limits = c("5%", "14%"),plot=FALSE) # assign q2
q2 <- qcc(dataset, type="xbar", nsigmas=1, labels=months, xlab= "Month", ylab = "Service Level %", title = "Phone Call SVL", digits=3, label.limits = c("76%", "80%"), plot=FALSE)
and then use each q1 & q2 with a plot method allowing to erase the y axis and redifine a new one with somekind of thicks.
plot(q1, yaxt="n", ... , xlab="Months", ylab="Service Levels %", title="Phone Call SVL", label.limits= c("5%", "13%"))
plot(q2, yaxt="n", ..., xlab="Months", ylab="Service Levels %", title="Phone Call SVL", label.limits= c("76%", "80%"))
Tom Hopper has written a blog on rewriting qcc plot using ggplot2 and grid,
though I haven't got a chance to dive deep into it. I wish I could bring you an easiy solution on y axis percentage formatting
though I haven't found one with qcc. Sorry.