Discrepancy between mlba::BostonHousing and code for creating Figure 3.1

[WACKE127]

In the BostonHousing.csv.gz file, there is a column "CAT. MEDV". The code references this value as "CAT.MEDV" but R coerces the whitespace when using read.csv() to be another '.', resulting in the column representing as "CAT..MEDV" to R. This causes downstream errors when using the aggregation method and attempting to reference this misnomer.

Problem can be resolved by either changing the CSV file's column name to be "CAT.MEDV" by deleting the whitespace, or by altering the source code for Figure 3.1 to be "CAT..MEDV"

#Non-functional Code
CAT.MEDV.per.CHAS <- aggregate(housing.df$CAT.MEDV,
by=list(housing.df$CHAS),
FUN=mean)
names(CAT.MEDV.per.CHAS) <- c("CHAS", "MeanCATMEDV")
CAT.MEDV.per.CHAS$CHAS <- factor(CAT.MEDV.per.CHAS$CHAS)

#Functional Code
CAT.MEDV.per.CHAS <- aggregate(housing.df$CAT..MEDV,
by=list(housing.df$CHAS),
FUN=mean)
names(CAT.MEDV.per.CHAS) <- c("CHAS", "MeanCATMEDV")
CAT.MEDV.per.CHAS$CHAS <- factor(CAT.MEDV.per.CHAS$CHAS)

[gedeck]

Thank you for your feedback. For some datasets, I preprocess the data before making them available through the package. For example, the Boston housing data are processed like this:

BostonHousing = read.csv(file.path('data-raw', 'BostonHousing.csv.gz')) %>%
  rename(CAT.MEDV=CAT..MEDV)
use_data(BostonHousing, overwrite=TRUE)

The code is in https://github.com/gedeck/mlba/blob/main/mlba/data-raw/0-prepareDataFiles.R. I do the renaming there and the processed file is available as mlba::BostonHousing If you use housing.df <- mlba::BostonHousing the code in the book works.

I left the original files unchanged, because we use them for other editions of the book as well.