Exploring the distribution of status and tree diameter
-Two columns that are commonly useful in ForestGEO datasets are status and dbh (diameter at breast height). We will begin by better understanding what type of variables they are. For this, base R provides useful functions.
status is a categorical variable.
We can count the number of observations in each category with table(), then visualize the result with barplot().
dbh is a continuous numeric variable.
summary(stem6$dbh)
-#> Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
-#> 10.50 33.25 127.50 143.63 188.00 528.00 8(Note the missing values (NAs).)
And we can visualize its distribution with hist().
Unfortunately hist() dropped missing values silently. But we can better understand how missing values of dbh relate to status by extracting only the columns dbh and status, and picking only the rows where dbh is missing.
dbh_status <- stem6[c("dbh", "status")]
-missing <- filter(dbh_status, is.na(dbh))
-unique(missing)
-#> # A tibble: 2 x 2
-#> dbh status
-#> <dbl> <chr>
-#> 1 NA D
-#> 2 NA GAnother approach is to count missing values.
-missing <- transform(stem6, na = ifelse(is.na(dbh), TRUE, FALSE))
-table(missing$na, missing$status)
-#>
-#> A D G
-#> FALSE 22 0 0
-#> TRUE 0 4 4We learn that dbh is missing where a tree is dead (status = D) or gone (status = G). This makes sense and, depending on the type of analysis we want to do, we may want to keep or remove missing values.
