R - Mean always returns 0 - r

Question

My goal is to create a function that reads specified .csv files (all of which have the same format) from the working directory, bind them into one data frame, and then return the mean of a specified column ("nitrate" or "sulfate") of that data frame. The current problem is that every time I call the function no matter how many files I chose to read/how many rows the mean is calculated on, the function always returns 0. I'm not quite sure how to fix this, any help appreciated.
pollutantmean <- function(pollutant, id = 1:332, directory =
"/Users/marsh/datasciencecoursera/specdata/") {
setwd(directory)
list <- list.files()
df <- data.frame()
for(i in id) {
x <- read.csv(list[i])
df <- rbind(df,x)
}
mean(!is.na(df["pollutant",]))
}

Answer 1

If you want the mean and there are NA's present in your data use: mean(df["pollutant",], na.rm = TRUE)
You are calculating the ratio of how many NA's are present. If you get 0 back, it means that you have only NA's . So maybe there is something else wrong as well. Maybe you can use dput() on a dataframe so we can have a look.

Answer 2

The syntax is wrong. It should be
mean(!is.na(df[[pollutant]]))
We don't need quotes for pollutant and should be same as the input argument. Secondly, the , implies that we are selecting the row names as the general format for indexing is row, column, which can be either numeric or character strings. In this case, we need to calculate the mean of missing values for a specific column. So [[ will extract the column or we can do
mean(!is.na(df[,pollutant]))
The whole function should now be
pollutantmean <- function(pollutant, id = 1:332, directory =
"/Users/marsh/datasciencecoursera/specdata/") {
setwd(directory)
list <- list.files()
df <- data.frame()
for(i in id) {
x <- read.csv(list[i])
df <- rbind(df,x)
}
mean(!is.na(df[[pollutant]]))
}
This can also be optimized using data.table
library(data.table)
pollutantmean <- function(pollutant, id = 1:332, directory =
"/Users/marsh/datasciencecoursera/specdata/") {
setwd(directory)
lst <- list.files()
df <- rbindlist(lapply(lst, fread))
mean(!is.na(df[[pollutant]]))
}

Answer 3

Just a guess, because there is no data to confirm this, but it looks like you are asking for the mean of the rows labeled pollutant, not the columns.
Typically a variable is saved in a column and individual observations are saved in rows. So, moving that comma will help get the right data into your calculation, giving you all rows(observations) of the column="pollutant".
#how the data frame is constructed df[rows, columns]
By asking the way that you did, you got all of the observations that do not have an NA in that row but you took the mean of the entire data frame.
pollutantmean <- function(pollutant, id = 1:332, directory =
"/Users/marsh/datasciencecoursera/specdata/") {
setwd(directory)
list <- list.files()
df <- data.frame()
for(i in id) {
x <- read.csv(list[i])
df <- rbind(df,x)
}
mean(df[,pollutant], rm.na=TRUE)
}
this says take the mean of all observations in the dataframe df column pollutant that are not = to NA this should give you what you want

Answer 4

All the above answers helped me fix it.
mean(df[[pollutant]], na.rm = TRUE)
ended up returning the correct answers. Thanks!