Merge branch 'main' of https://github.com/UGA-IDD/SISMID-2024

exercises/Exercise03-Soln.R

@@ -3,3 +3,159 @@
 # Exercise 03 Solution
 # Zane Billings and Amy Winter
 ################################################################################
+
+# In this exercise, we'll practice data cleaning using the measles datasets
+# we've looked at previously.
+
+# Question 1 ###################################################################
+# Part A: Load in the MeaslesCases.csv data, and take a look at it to remind
+# yourself of the structure.
+cases <- read.csv(here::here("data", "MeaslesCases.csv"))
+str(cases)
+
+# Part B: Is this data in wide format or long format? Explain your answer.
+
+# The data is in wide format because there are multiple columns that store
+# the number of measles cases. Each year of data is a column, so the same
+# variable is stored across multiple columns.
+
+# Part C: use the reshape() function to edit this dataset so that you have
+# one column for cases and one column for year.
+cases_long <-
+	reshape(
+		cases,
+		direction = "long",
+		varying = which(startsWith(names(cases), "X")),
+		v.names = "Cases",
+		idvar = names(cases)[1:3],
+		times = 1980:2023
+	)
+
+# Part D: look at the str() of the long format data. What is new compared to
+# what we usually see? Try calling reshape() again with only your long format
+# data set as an argument.
+str(cases_long)
+reshape(cases_long)
+
+# This data has an attr() field which adds additional metadata to it. This can
+# be used by reshape() if we need to pivot the data back and forth.
+
+# Part E: remove the description column from this data set.
+cases_long$Description <- NULL
+
+# Question 2 ###################################################################
+# Part A: Load in the PopSize data using an appropriate package.
+library(readxl)
+popsize <- readxl::read_excel(here::here("data", "PopSize.xlsx"))
+
+# Part B: Remove the "indicator name" and "indicator code" variables.
+popsize$`Indicator Name` <- NULL
+popsize$`Indicator Code` <- NULL
+
+# Part C: reshape the dataset into long format.
+# Make sure the times in the long format data are correct and the variable
+# containing the population size data has an understandable name.
+popsize_long <-
+	reshape(
+		popsize,
+		direction = "long",
+		varying = names(popsize)[3:66],
+		v.names = "total_pop",
+		idvar = names(popsize)[1:2],
+		times = 1960:2023
+	)
+
+# Part D: rename any variable names that have spaces in them.
+names(popsize_long)[names(popsize_long) == "Country Name"] <- "country_name"
+names(popsize_long)[names(popsize_long) == "Country Code"] <- "country_code"
+
+# Part E: Finally, Look at the str() of your new data frame at the end to
+# double check everything.
+str(popsize_long)
+
+# Question 3 ###################################################################
+# Part A: Read in the MeaslesVaccinationCoverage.csv data set.
+vc <- read.csv(here::here("data", "MeaslesVaccinationCoverage.csv"))
+
+# Part B:
+# Merge together the (long format) cases dataset and the vaccination coverage
+# dataset. Set the `all` argument to be TRUE (i.e., do a full join).
+# You will need to specify the "by.x" and "by.y" arguments correctly.
+cases_vc <-
+	merge(
+		cases_long, vc,
+		by.x = c("iso3c", "time"),
+		by.y = c("CODE", "YEAR"),
+		all = TRUE
+	)
+
+# Part C:
+# Merge the previous cases/coverage dataset with the long form of the popsize
+# dataset. Make sure you include all years/countries with records in the
+# cases/coverage dataset but do NOT include years/countries from the popsize
+# dataset that are unused. (I.e. if x = cases/coverage dataset, all.x = TRUE
+# and all.y = FALSE).
+cases_vc_pop <-
+	merge(
+		cases_vc, popsize_long,
+		by.x = c("iso3c", "time"),
+		by.y = c("country_code", "time"),
+		all.x = TRUE
+	)
+
+# Part D: notice that you now have three variables that contain country names.
+# Keep the variable "Country...Region" and get rid of the other two. Then,
+# rename that variable to "country".
+# Then, remove the "ANTIGEN_DESCRIPTION" column, and rename any columns that
+# are all uppercase to have better formatted names.
+cases_vc_pop$country_name <- NULL
+cases_vc_pop$NAME <- NULL
+names(cases_vc_pop)[names(cases_vc_pop) == "Country...Region"] <- "country"
+
+cases_vc_pop$ANTIGEN_DESCRIPTION <- NULL
+names(cases_vc_pop)[names(cases_vc_pop) == "ANTIGEN"] <- "vaccine_antigen"
+names(cases_vc_pop)[names(cases_vc_pop) == "COVERAGE"] <- "vaccine_coverage"
+
+# Part E: sort the dataset by country code, then by year, then by the variable
+# containig the MCV1/MCV2 information (whatever you renamed it).
+# Hint: use the order() function.
+sort_order <-
+	order(
+		cases_vc_pop$iso3c, cases_vc_pop$time, cases_vc_pop$vaccine_antigen
+	)
+meas_sorted <- cases_vc_pop[sort_order, ]
+
+# Part F: remove all of the countries where the MCV1/MCV2 variable is missing.
+meas_clean <- meas_sorted[!is.na(meas_sorted$vaccine_antigen), ]
+
+# Part G: if you look at all of the country names in the dataset, you will see
+# that some are NA.
+
+# Part G: use some of the functions we've learned to look at your dataset
+# and make sure it looks OK. Then answer these questions.
+# - How many rows are there in the final data set?
+# - How many different countries are there in the final data set?
+# - What year did MCV2 first start being reported in the data? (Hint:
+#    use the table() function.)
+# - How many countries have observations in every year of the data set? (Hint:
+#    use the table() function, and then use rowSums() on the table.)
+
+nrow(meas_clean)
+length(unique(meas_clean$country))
+table(meas_clean$time, meas_clean$vaccine_antigen) # The answer is 2000
+
+year_counts_per_country <-
+	table(meas_clean$country, meas_clean$time) |>
+	rowSums()
+
+max_years <- max(year_counts_per_country)
+
+names(year_counts_per_country)[year_counts_per_country == max_years] |>
+	length()
+
+# Part F: save your cleaned dataset to a .Rds file in your data directory!
+# Name it "measles_final.Rds".
+saveRDS(meas_clean, here::here("data", "measles_final.Rds"))
+
+#### Congratulations on finishing exercise 3! This one was tough but look at
+# all the cool and useful stuff you've already learned to do!