tests/testthat/test_kmeans_2.R

In exploratory-io/exploratory_func: R package for Exploratory

# how to run this test:
# devtools::test(filter="kmeans_2")

context("test kmeans analytics view functions with California Election data.")

df <- exploratory::read_delim_file("https://www.dropbox.com/s/5dhy5hcq1b8zla4/CA_Election_spreaded.csv?dl=1" , ",", quote = "\"", skip = 0 , col_names = TRUE , na = c('','NA') , locale=readr::locale(encoding = "UTF-8", decimal_mark = ".", grouping_mark = "," ), trim_ws = TRUE , progress = FALSE) %>%
  readr::type_convert() %>%
  exploratory::clean_data_frame()

test_that("exp_kmeans result is properly separated out.", {
  df[[3]][[2]] <- NA # Inject NA for testing NA filtering.
  df <- df %>% mutate(all_na=NA) # Add all-NA row to test NA handling.
  model_df <- df %>% exp_kmeans(all_na, `Adult Film Condom Requirements`, `Ban on Single-use Plastic Bags`, `Carryout Bag Charges`, `Cigarette Tax`, `Corporate Political Spending Advisory Question`, `Criminal Sentences & Juvenile Crime Proceedings`, `Death Penalty Procedure Time Limits`, `English Proficiency. Multilingual Education.`, `Firearms and Ammunition Sales`, `K-12 and Community College Facilities`, `Legislative Procedure Requirements`, `Marijuana Legalization`, `Medi-Cal Hospital Fee Program`, `Repeal of Death Penalty`, `State Prescription Drug Purchase Standards`, `Tax Extension for Education and Healthcare`, `Voter Approval of Revenue Bonds`, algorithm = "Hartigan-Wong", normalize_data = TRUE, elbow_method_mode = FALSE, centers=2)
  res <- model_df %>% tidy_rowwise(model, type="data")
  # Verify that on PC1 axis, 2 cluster means are separated more than the sum of each group's standard deviation.
  summarized <- res %>% group_by(cluster) %>% summarize(m=mean(PC1),sd=sd(PC1))
  expect_true(abs(summarized$m[[1]] - summarized$m[[2]]) > summarized$sd[[1]] + summarized$sd[[2]])
})

test_that("exp_kmeans result is properly separated out.", {
  df[[3]][[2]] <- NA # Inject NA for testing NA filtering.
  model_df <- df %>% build_kmeans.cols(`Adult Film Condom Requirements`, `Ban on Single-use Plastic Bags`, `Carryout Bag Charges`, `Cigarette Tax`, `Corporate Political Spending Advisory Question`, `Criminal Sentences & Juvenile Crime Proceedings`, `Death Penalty Procedure Time Limits`, `English Proficiency. Multilingual Education.`, `Firearms and Ammunition Sales`, `K-12 and Community College Facilities`, `Legislative Procedure Requirements`, `Marijuana Legalization`, `Medi-Cal Hospital Fee Program`, `Repeal of Death Penalty`, `State Prescription Drug Purchase Standards`, `Tax Extension for Education and Healthcare`, `Voter Approval of Revenue Bonds`, centers=2, keep.source=TRUE, augment = FALSE)
  res <- model_df %>% augment_kmeans(model, data=source.data)
  # Verify that on Cigarette.Tax values, 2 cluster means are separated more than the sum of each group's standard deviation.
  summarized <- res %>% group_by(cluster) %>% summarize(m=mean(`Cigarette Tax`),sd=sd(`Cigarette Tax`))
  expect_true(abs(summarized$m[[1]] - summarized$m[[2]]) > summarized$sd[[1]] + summarized$sd[[2]])
})