Trying to Add Loess Smoothing Curve to Scatterplot

Question

I am trying to add a loess smoothed fit curve to my scatterplot in R. I can't seem to figure out what's wrong with my code below...For reference, the variables poverty and binge_all are column names of a the data frame correlational_data. I have loaded the ggplot2 package/library.

library(ggplot2)    

p <- ggplot(correlational_data, aes(poverty, binge_all))
p <- p + geom_point(color = "blue")
p <- p + geom_smooth(method = "loess")
p

I used sapply(correlational_data$poverty, class) and sapply(correlational_data$binge_all, class) to determine that poverty and binge_all are of class factor. Not sure if that makes a difference.

Update to show first 10 rows of data

head(correlational_data, 10)
   year                state binge_all poverty
1  2012              Alabama      12.3      19
2  2012               Alaska      16.8    10.1
3  2012              Arizona      15.3    18.7
4  2012             Arkansas      11.8    19.8
5  2012           California      16.9      17
6  2012             Colorado      19.2    13.7
7  2012          Connecticut      17.5    10.7
8  2012             Delaware      18.6      12
9  2012 District of Columbia      23.1    18.2
10 2012              Florida      16.5    17.1

poverty and binge_all need to be numeric for geom_smooth to work. It will be easier to help you further if you provide all of the relevant code you ran plus a sample of your data that runs with your code (for example, paste into your question the output of dput(correlational_data[1:10, ]) to provide the first ten rows of your data). — eipi10
To check the class of a single column: class(correlational_data$poverty). sapply(correlational_data, class) will return the class of every column in the data frame. — eipi10
@eipi10 does the update that I posted help to answer your question? — PugFanatic
No; show the result of str(correlational_data) and read @epi s comment again — user101089

www www · Accepted Answer · 2017-09-26T01:58:47

As others pointed out in the comments, binge_all and poverty needs to be numeric, not factor. Here I plot the data using the code and example data you provided.

# Create example data frame
correlational_data <- read.table(text = "   year                state binge_all poverty
1  2012              Alabama      12.3      19
                                 2  2012               Alaska      16.8    10.1
                                 3  2012              Arizona      15.3    18.7
                                 4  2012             Arkansas      11.8    19.8
                                 5  2012           California      16.9      17
                                 6  2012             Colorado      19.2    13.7
                                 7  2012          Connecticut      17.5    10.7
                                 8  2012             Delaware      18.6      12
                                 9  2012 'District of Columbia'      23.1    18.2
                                 10 2012              Florida      16.5    17.1",
                                 header = TRUE, stringsAsFactors = FALSE)

# Check the class
class(correlational_data$binge_all)
[1] "numeric"
class(correlational_data$poverty)
[1] "numeric"

# Plot the data   
library(ggplot2)

p <- ggplot(correlational_data, aes(poverty, binge_all))
p <- p + geom_point(color = "blue")
p <- p + geom_smooth(method = "loess")
p

Notice that if you want to convert your factor column to numeric, please convert to character first. Below is an example:

correlational_data$binge_all <- as.numeric(as.character(correlational_data$binge_all))
correlational_data$poverty <- as.numeric(as.character(correlational_data$poverty))

This will make sure you convert the actual numbers, not the level of the factor.

Trying to Add Loess Smoothing Curve to Scatterplot

1 Answers