sars2pack_workflow_gb_2020-06-16.R

browseURL('https://seandavi.github.io/ITR/r_intro_mechanics.html')
1 + 1
x = 10
x
y = 20
y
y + x

browseURL('https://seandavi.github.io/ITR/vectors.html')
# vectors
b = 1:10
b

# subset
b[3]
b[7]
b[3:7]

# data.frame
browseURL('https://seandavi.github.io/ITR/dataframes_intro.html')
d = data.frame(x=1:10,y=letters[1:10])
d
head(d)

# subset uses [] square brackets with [row, column]
d[1,2]
d[2,1]
d[1,]
d[2:5,]



# browseURL('https://seandavi.github.io/ITR/control_looping_functions.html')
# make a function
add_x_and_y = function(xval, yval) {
    return(xval + yval)
}

# Use a function
# functions have () after them
add_x_and_y(20,100)


browseURL('https://seandavi.github.io/ITR/dplyr_intro_msleep.html')
library(dplyr) 
library(purrr)
library(ggplot2)
browseURL('https://seandavi.github.io/sars2pack/')
library(sars2pack)

# Work with United States disease stats by state
# I know which dataset to use, but if you don't
# browseURL('https://seandavi.github.io/sars2pack/articles/datasets.html')

help('nytimes_state_data')

nyt = nytimes_state_data()

head(nyt)
table(nyt$subset)
table(nyt$state)
summary(nyt$date)

nyt_ny = nyt %>% dplyr::filter(state=='New York')
table(nyt_ny$state)

nyt_ny_confirmed = nyt_ny %>% dplyr::filter(subset == 'confirmed')
table(nyt_ny_confirmed$subset)

help('plot_epicurve')
plot_epicurve(nyt_ny_confirmed, date_column = 'date', 
              case_column = 'count')

######################
# Challenge 1
######################
# Change the code above to look at a different
# state.
# Bonus: Change the code above to look at multiple
# states.

# Have been looking at cumulative cases
# How about incidence?
head(nyt_ny_confirmed)
nyt_ny_incidence = nyt_ny_confirmed %>% 
    add_incidence_column()
head(nyt_ny_incidence)

plot_epicurve(nyt_ny_incidence, date_column = 'date', 
              case_column = 'inc', log=FALSE)

plot_epicurve(nyt_ny_incidence, date_column = 'date', 
              case_column = 'inc', log=TRUE)

# Bonus: mark day of week (green is Sunday)
plot_epicurve(nyt_ny_incidence %>% 
                  mutate(dow = lubridate::wday(date)), 
              date_column = 'date', 
              case_column = 'inc', log=TRUE) + 
    geom_point(aes(col=dow==1))


# When did New York hit its peak in cases?
nyt_ny_incidence[which.max(nyt_ny_incidence$inc),]

# What are the counts over the last week?
nyt_ny_incidence %>% dplyr::filter(date>=Sys.Date()-7)