Easy to make read_csv typo results in different data

[ebolyen]

This is really just a note for the future.

At STAMPS we were teaching section 3 of this lesson, and some people were getting a different number of rows after this command.

This is a result of a simple typo using read.csv instead of read_csv which results in a dataframe with a different number of rows.

[ryanpeek]

Perhaps we could create a dataset where whether we use read.csv or read_csv we get the same solution...particularly because it is a fairly easy fix and it can really sidetrack instructors/students from getting at the meat of the lesson. I'll try to work on a PR that does this.

[tracykteal]

Thanks @ryanpeek. If we could fix those entries in the dataset, that would be good and help with that confusion for instructors and students. It would be good to keep the overall number of rows the same because those are referenced in other parts of the lesson. We would need to update the dataset on Figshare, but we can do that.

[ryanpeek]

Hi folks (@tracykteal @fmichonneau) ,
So attached is the "revised" portals dataset...I've vetted it with both the data lessons (02-starting-with-data.Rmd & 03-dplyr.Rmd. I replaced the blanks in the surveys$sex column with "blank", which then explicitly reads in as a character or a factor using either read_csv/read.csv. All rows/cols remain the same, and the section on factors with the barplots still works, since it's indexed by number and not by name.

The only thing that needs to be updated is the barplot figure in the "starting with data lesson". I'm not sure how the overall R-ecology site is built/knitted, but if someone reknits the Rmd it should take care of itself (no code changes needed).

The code I used to make this change and write it out is below, let me know if you all need anything else! :)

download.file(url="https://ndownloader.figshare.com/files/2292169",
              # destfile = "data/portal_data_joined.csv")
surveys.csv <- read.csv("data/portal_data_joined.csv")

# Change the "" in the sex to a character level called "blank". This would still permit using barplot in the starting with data lesson
library(dplyr)
surveys.csv.revised <- surveys.csv %>% 
  mutate(
    sex = as.character(sex),
    sex = case_when(
    grepl("^$", sex) ~ "blank",
    TRUE ~ sex))

str(surveys.csv.revised$sex)
summary(as.factor(surveys.csv.revised$sex))

write.csv(surveys.csv.revised, "data/portal_data_joined_revised.csv", row.names = FALSE, na = "")

portal_data_joined_revised.csv.txt

[fmichonneau]

addressed with #663