Logarithmic scaled scatter chart?

Can you check the data type that you're using for the x-axis? I'm using whole numbers in the example below and seeing the option to change the scale from linear to log. If you're not able to get it to work, one other workaround would be to create an additional column with the log value, then use that in the scatter plot.

Thanks,
Sam Lester (MSFT)

from what I can see the base graphs in Power BI aren't particularly flexible, but that's fine, because they've included support for R.

This means you can use the ggplot2 library, and so can customise pretty much everything you could ever want to (and quite a few things you wouldn't).

What you need to do

Click on the R symbol under visualisations

Select the fields you want to chart

paste this code into the R script editor section

library(ggplot2)
names(dataset) <- gsub(" ","",names(dataset)) #only needed if you have spaces in field names

g <- ggplot(dataset,aes(x=PUT_X_FIELD_HERE,y=PUT_Y_FIELD_HERE))
g <- g + geom_point(color="#D0021A") #change the #D0021A to whatever colour you want
g <- g + scale_x_log10(label=function(x){return(x))} #you could change this bit if you wanted to display (for example) the log values on the axis instead of the raw values
 ,limits=c(which.min(dataset$PUT_X_FIELD_HERE>0),max(dataset$PUT_X_FIELD_HERE)) #this is the bit that actually does the work
g <- g + scale_y_log10(label=function(y){return(y))}
 ,limits=c(which.min(dataset$PUT_X_FIELD_HERE>0),max(dataset$PUT_Y_FIELD_HERE))
g <- g + coord_equal() #this forces the graph to be square. Delete if that's not what you want
g <- g + ggtitle("PUT GRAPH TITLE HERE")+labs(x="X AXIS LABEL GOES HERE",y="Y AXIS LABEL GOES HERE")
g <- g + theme(plot.title = element_text(size=20,face="bold",margin=margin(10,0,10,0)))

g

replacing PUT_X_FIELD_HERE with whatever your x field is called, without spaces, and likewise for y.

This might seem like taking a sledgehammer to crack a walnut, but you'll find that doing anything non-standard is significantly less hassle in R in the long run.

If you need any details on what the elements of ggplot actually do, I found this cheatsheet really helpful when I was learning this.

clarification:

g <- g + scale_x_log10(label=function(x){return(x))},limits=c(which.min(dataset$PUT_X_FIELD_HERE>0),max(dataset$PUT_X_FIELD_HERE))

this bit is doing the actual work here, and is basically saying  "make x a log scale between the lowest positive x value (to avoid the log(0) issue) and the highest value"