# Title: Lab 1
# Author: Justin Rasmussen
# Date created: 2021-01-25
# Date edited: 2021-01-25
# Description: Measures of central tendency, basic data manipulation & visualization

# Load packages
library(tidyverse)
library(lubridate)

# Import data

# Import data from NYT github
nyt_st_url <- "https://raw.githubusercontent.com/nytimes/covid-19-data/master/us-counties.csv"
nyt_cnt <- read.csv(nyt_st_url, stringsAsFactors = FALSE)
# Export locally to avoid slow internet issues
write.csv(nyt_cnt, "nyt_cnt.csv")


# Basic data manipulation
#==============================================================================+

# The dplyr package uses a few important verbs to manipulate data:

# Select
#----------
# this lets you choose variables (columns) in your dataframe
nyt_cnt_select <- select(nyt_cnt, date, county, state, cases, deaths)
# sometimes it's more efficient to just drop the ones you don't want
nyt_cnt_select <- select(nyt_cnt, -fips)

# Mutate
#--------
# mutate lets you create new variables easily
nyt_cnt_mutate <- mutate(nyt_cnt_select,
                         date = ymd(date))
# quick note here: logical operators in R are:
# equal to:                      ==
# not equal to:                  !=
# greater than:                  >
# greater than or equal to:      >=
# less than:                     <
# less than or equal to:         <=
# not:                           !
# or:                            |
# and:                           &

# Filter
#--------
# filter lets you choose observations (rows) in your dataframe
nyt_cnt_filter <- filter(nyt_cnt_mutate,
                         date > ymd("2020-12-01"),
                         state == "North Carolina")

# Summarize
#----------
# summarize lets you create aggregate tables
nyt_cnt_summarize <- summarize(nyt_cnt_filter,
                               mean_cases = mean(cases),
                               median_cases = median(cases),
                               min_cases = min(cases),
                               max_cases = max(cases))

# Group_by
#----------
# group_by() lets you create grouped tables to summarize
nyt_cnt_grouped <- group_by(nyt_cnt_filter, state_county)
nyt_cnt_summarize <- summarize(nyt_cnt_grouped,
                               mean_cases = mean(cases),
                               median_cases = median(cases),
                               min_cases = min(cases),
                               max_cases = max(cases))

# Piping
#--------
# this IS a pipe: %>%
# a pipe allows you to pass the result of one function
group_by(nyt_cnt_filter, county) %>%
  # on to the data argument of the next function
  summarize(mean_cases = mean(cases),
            median_cases = median(cases),
            min_cases = min(cases),
            max_cases = max(cases))
# you can also use . to represent the data being piped in: 
nyt_cnt_filter %>% group_by(., county) %>%
  summarize(.,
            mean_cases = mean(cases),
            median_cases = median(cases),
            min_cases = min(cases),
            max_cases = max(cases))

# with a pipe we can perform data manipulation wihtout intermediate objects
nyt_cnt_summarize <- nyt_cnt %>%
  select(-fips) %>%
  mutate(date = ymd(date)) %>%
  filter(date > ymd("2020-12-01"),
         state == "North Carolina") %>%
  group_by(county) %>%
  summarize(mean_cases = mean(cases),
            median_cases = median(cases),
            min_cases = min(cases),
            max_cases = max(cases),
            mean_deaths = deaths)

# Basic data visualization
#===============================================================================
# ggplot2 is a package that builds plots in layers
nyt_cnt_filter %>%
  # mapping data to aesthetics
  ggplot(aes(y = cases, x = date, color = county)) +
  # choosing a geometry to represent those aesthetics
  geom_line() +
  # lots of other layers to play with e.g.
  theme_minimal()

# you can click "Export" directly from the plot window...
# or "Zoom" and then right-click to "Save as picture"

# Challenge
#===============================================================================
# Calculate the mean cases and deaths for 3 states of your choice in 2021
# Create a graph showing how deaths and cases have changed in each state