mapStats (version 3.1)

jiggleClass: Adjust class boundaries to protect from rounding errors

Description

When using classIntervals to compute classes, occasionally there are rounding errors so that when the data is plotted and the class breaks are used for colors, for instance, the rounding error may cause a value to not be plotted with the right color, or to not be plotted at all. For this reason, we add a small value to each of the break points to accomodate a possible rounding error. This correction is negligible and should not affect plotting. When both the variable values and all the interval breaks are integer-valued (e.g. for sums of integer values), jiggleClass adjusts the interval breaks to be at the midpoints between interval endpoints. Additionally, in case ngroups is high, resulting in empty groups (even though the number of unique values is higher than ngroups), the function also eliminates the empty groups as part of the adjustment above. In case there is such a change, the palettes are also changed.

Usage

jiggleClass(x)

Value

an object of class classIntervals.

Arguments

x

an object of class classIntervals from the function classIntervals.

Examples

y <- 100*rnorm(50)

#compute class intervals using either right or left interval closure

class_list_right <- classInt::classIntervals(var=y, n=12, intervalClosure="right")
class_list_right$brks
class_list_left <- classInt::classIntervals(var=y, n=12, intervalClosure="left")
class_list_left$brks

#there should be a slight difference now between class breaks from before, but should 
#have same number of observations per interval as before, and for both left and right closure

jiggleClass(x=class_list_right)
jiggleClass(x=class_list_left)


#example with discrete data, 7 groups but 9 unique values.
#classIntervals generates some empty intervals, so jiggleClass eliminates them and adjusts breaks
#in this example, with integer values, right/left closure matters more, and so the results
#will differ slightly depending on which is chosen
# here, both the observed values and the breaks under right and left are integers

y <- c(1, 1, 1, 1, 2, 3, 6, 7, 8, 9, 10, 10, 10, 10, 11)
class_list_right <- classInt::classIntervals(y, 7, intervalClosure="right")
class_list_right
#style: quantile
#  one of 28 possible partitions of this variable into 7 classes
#  [1,1]   (1,2]   (2,6]   (6,8]  (8,10] (10,10] (10,11] 
#      4       1       2       2       1       4       1 

class_list_left <- classInt::classIntervals(y, 7, intervalClosure="left")
class_list_left
#style: quantile
#  one of 28 possible partitions of this variable into 7 classes
#  [1,1)   [1,2)   [2,6)   [6,8)  [8,10) [10,10) [10,11] 
#      0       4       2       2       2       0       5 


#number of groups falls now for left closure, from 7 to 5, in this example

jiggleClass(x=class_list_right)
#style: quantile
#  one of 28 possible partitions of this variable into 7 classes
#  [0.5,1.5]   (1.5,2.5]   (2.5,6.5]   (6.5,8.5]   (8.5,9.5]  (9.5,10.5] (10.5,11.5] 
#          4           1           2           2           1           4           1 


jiggleClass(x=class_list_left)
#style: quantile
#  one of 70 possible partitions of this variable into 5 classes
# [0.5,1.5)  [1.5,4.5)  [4.5,7.5)  [7.5,9.5) [9.5,11.5] 
#         4          2          2          2          5