plot_data_for_review.Rmd

---
title: "Data for Review"
output: html_document
---

```{r echo=FALSE}
library(tidyverse)
library(covidData)
library(plotly)
library("googlesheets4")

# do we want to plot historical data available as of all past weeks or only the
# most recent?
plot_all_historical_as_ofs <- FALSE

# plot historical records of cases and deaths using
# data available as of Sunday each week.
most_recent_sunday <- lubridate::floor_date(Sys.Date(), unit = "week")

# # # # use fixed date
# most_recent_sunday<-  as.Date("2021-05-16")
previous_sunday<-most_recent_sunday - 7

if (plot_all_historical_as_ofs) {
  first_as_of_dates <- data.frame(
    measure = c("cases", "hospitalizations", "deaths"),
    first_as_of_date = as.Date(c("2020-01-04", "2020-01-04", "2020-01-04"))
  )
} else {
  first_as_of_dates <- data.frame(
    measure = c("cases", "hospitalizations", "deaths"),
    first_as_of_date = rep(most_recent_sunday, 3)
  )
}

# date ranges for plots
plot_start_date <- lubridate::ymd("2020-01-01")
plot_end_date <- most_recent_sunday

# locations to plot: state and national have 2 digit location codes
all_locations <- covidData::fips_codes %>%
  dplyr::filter(nchar(location) == 2) %>%
  dplyr::pull(location)
```

# Plots by Target Variable {.tabset .tabset-fade}

## Cases

```{r fig.height=8, fig.width=12, results='asis', echo=FALSE}
# read in outliers
case_outliers <- read_sheet(
  ss = "https://docs.google.com/spreadsheets/d/1Vw1Oakr-KdLB8RJoZNF7u6MRXMzg6iSJLpT_PpaKi2Y/edit#gid=799534143",
  range = "cases"
) %>%
  dplyr::mutate(
    date = as.Date(date),
    issue_date = as.Date(issue_date)
  ) %>%
  filter (issue_date == previous_sunday, reported_inc > 0) %>%
  select (location, date, reported_inc)

# make the plots
measure <- "cases"

# vector of all as_of dates to plot
all_as_ofs <- seq.Date(
  from = first_as_of_dates %>%
    dplyr::filter(measure == UQ(measure)) %>%
    dplyr::pull(first_as_of_date),
  to = most_recent_sunday,
  by = 7
)

# temporal resolution depends on measure
if (measure == "hospitalizations") {
  temporal_resolution <- "daily"
} else {
  temporal_resolution <- "weekly"
}

# collect all required data up front to reduce number of calls to load_data
data_all_locations <- purrr::map_dfr(
  all_as_ofs, # used for as_of argument to covidData::load_data
  function(as_of) {
    covidData::load_data(
      as_of = as_of,
      spatial_resolution = c("state", "national"),
      temporal_resolution = temporal_resolution,
      measure = measure
    ) %>%
      dplyr::mutate(
        as_of = as_of
      )

  }
)

output <- list()
for (location in all_locations) {
  location_name <- covidData::fips_codes %>%
    dplyr::filter(location == UQ(location)) %>%
    dplyr::pull(location_name)

    abbreviation <- covidData::fips_codes %>%
    dplyr::filter(location == UQ(location)) %>%
    dplyr::pull(abbreviation)
    
  for (as_of in as.character(all_as_ofs)) {
#    message(paste0(location_name, ", ", as_of))
    data_to_plot <- data_all_locations %>%
      dplyr::filter(
        location == UQ(location),
        as_of == UQ(as_of),
        date >= plot_start_date)

    # subset negative points
        data_to_plot_neg <- data_to_plot %>%
        filter(inc < 0)
    
        # outliers to plot
      outlier_to_plot <- case_outliers %>%
      dplyr::filter(
        location == UQ(location),
        date >= plot_start_date)
            
    if (nrow(data_to_plot) > 0) {
      p <- ggplot() +
        geom_line(data = data_to_plot,
          mapping = aes(x = date, y = inc)) +
        geom_point(data = data_to_plot, aes(date, inc))+
        geom_point(data = data_to_plot_neg, aes(date, inc), color = 'red')+
        geom_point(data = outlier_to_plot, aes(date, reported_inc), color = 'orange')+
        scale_x_date(
          breaks = data_to_plot %>%
            dplyr::filter(weekdays(date) == "Saturday") %>%
            dplyr::pull(date) %>%
            unique(),
          minor_breaks = NULL,
          limits = c(plot_start_date, plot_end_date)) +
         ggtitle(paste0(location_name," (", abbreviation,") as of ", as_of)) +
        theme_bw() +
        theme(
          axis.text.x = element_text(angle = 90, vjust = 0.5),
          panel.grid.major.x = element_line(colour = "darkgrey")
        )
     
      output[[length(output) + 1L]] <- '#### \n'
      
      pp <- plotly::ggplotly(p)
      output[[length(output) + 1L]] <- pp
      output[[length(output) + 1L]] <- cat('\n \n')
    }
  }
}

# Render the outputs
for(j in 1:length(output)){
  x <- output[[j]]

  if(inherits(x, "character")){
    cat("\n")
    cat(x)
  } else if(inherits(x, "knitr_kable")){
    cat("\n")
    print(x)
  }
  else {
    # print the html piece of the htmlwidgets
    cat("\n")
    cat(htmltools::renderTags(as_widget(x))$html)
  }
}
```



## Hospitalizations

```{r fig.height=8, fig.width=12, results='asis', echo=FALSE}
# # read in outliers
# hosp_outliers <- read_sheet(
#   ss = "https://docs.google.com/spreadsheets/d/1Vw1Oakr-KdLB8RJoZNF7u6MRXMzg6iSJLpT_PpaKi2Y/edit#gid=799534143",
#   range = "hosps"
# ) %>%
#   dplyr::mutate(
#     date = as.Date(date),
#     issue_date = as.Date(issue_date)
#   ) %>%
# filter (issue_date == previous_sunday, reported_inc > 0) %>%
#   select (location, date, reported_inc)
# 
# # make the plots
# measure <- "hospitalizations"
# 
# # vector of all as_of dates to plot
# all_as_ofs <- seq.Date(
#   from = first_as_of_dates %>%
#     dplyr::filter(measure == UQ(measure)) %>%
#     dplyr::pull(first_as_of_date),
#   to = most_recent_sunday,
#   by = 7
# )
# 
# # temporal resolution depends on measure
# if (measure == "hospitalizations") {
#   temporal_resolution <- "daily"
# } else {
#   temporal_resolution <- "weekly"
# }
# 
# # collect all required data up front to reduce number of calls to load_data
# data_all_locations <- purrr::map_dfr(
#   all_as_ofs, # used for as_of argument to covidData::load_data
#   function(as_of) {
#     covidData::load_data(
#       as_of = as_of,
#       spatial_resolution = c("state", "national"),
#       temporal_resolution = temporal_resolution,
#       measure = measure
#     ) %>%
#       dplyr::mutate(
#         as_of = as_of
#       )
#   }
# )
# 
# output <- list()
# for (location in all_locations) {
#   location_name <- covidData::fips_codes %>%
#     dplyr::filter(location == UQ(location)) %>%
#     dplyr::pull(location_name)
  # abbreviation <- covidData::fips_codes %>%
  #   dplyr::filter(location == UQ(location)) %>%
  #   dplyr::pull(abbreviation)
# 
#   for (as_of in as.character(all_as_ofs)) {
# #    message(paste0(location_name, ", ", as_of))
#     data_to_plot <- data_all_locations %>%
#       dplyr::filter(
#         location == UQ(location),
#         as_of == UQ(as_of),
#         date >= plot_start_date)
# 
#  # subset negative points
#         data_to_plot_neg <- data_to_plot %>%
#         filter(inc < 0)
#         
#    # outliers to plot
#       outlier_to_plot <- hosp_outliers %>%
#       dplyr::filter(
#         location == UQ(location),
#         date >= plot_start_date)
# 
#     if (nrow(data_to_plot) > 0) {
#       p <- ggplot() +
#         geom_line(data = data_to_plot,
#           mapping = aes(x = date, y = inc)) +
#         geom_point(data = data_to_plot, aes(date, inc))+
#         geom_point(data = data_to_plot_neg, aes(date, inc), color = 'red')+
#         geom_point(data = outlier_to_plot, aes(date, reported_inc), color = 'orange')+
#         scale_x_date(
#           breaks = data_to_plot %>%
#             dplyr::filter(weekdays(date) == "Saturday") %>%
#             dplyr::pull(date) %>%
#             unique(),
#           minor_breaks = NULL,
#           limits = c(plot_start_date, plot_end_date)) +
#        ggtitle(paste0(location_name," (", abbreviation,") as of ", as_of)) +
#         theme_bw() +
#         theme(
#           axis.text.x = element_text(angle = 90, vjust = 0.5),
#           panel.grid.major.x = element_line(colour = "darkgrey")
#         )
#       # plotly::ggplotly(p) Need help
#       output[[length(output) + 1L]] <- '#### \n'
# 
#       pp <- plotly::ggplotly(p)
#       output[[length(output) + 1L]] <- pp
#       output[[length(output) + 1L]] <- cat('\n \n')
#     }
#   }
# }
# 
# # Render the outputs
# for(j in 1:length(output)){
#   x <- output[[j]]
# 
#   if(inherits(x, "character")){
#     cat("\n")
#     cat(x)
#   } else if(inherits(x, "knitr_kable")){
#     cat("\n")
#     print(x)
#   }
#   else {
#     # print the html piece of the htmlwidgets
#     cat("\n")
#     cat(htmltools::renderTags(as_widget(x))$html)
#   }
# }
```



## Deaths

```{r fig.height=8, fig.width=12, results='asis', echo=FALSE}
# read in outliers
death_outliers <- read_sheet(
  ss = "https://docs.google.com/spreadsheets/d/1Vw1Oakr-KdLB8RJoZNF7u6MRXMzg6iSJLpT_PpaKi2Y/edit#gid=799534143",
  range = "deaths"
) %>%
  dplyr::mutate(
    date = as.Date(date),
    issue_date = as.Date(issue_date)
  ) %>%
  filter (issue_date == previous_sunday, reported_inc > 0) %>%
  select (location, date, reported_inc)

# make the plots
measure <- "deaths"

# vector of all as_of dates to plot
all_as_ofs <- seq.Date(
  from = first_as_of_dates %>%
    dplyr::filter(measure == UQ(measure)) %>%
    dplyr::pull(first_as_of_date),
  to = most_recent_sunday,
  by = 7
)

# temporal resolution depends on measure
if (measure == "hospitalizations") {
  temporal_resolution <- "daily"
} else {
  temporal_resolution <- "weekly"
}

# collect all required data up front to reduce number of calls to load_data
data_all_locations <- purrr::map_dfr(
  all_as_ofs, # used for as_of argument to covidData::load_data
  function(as_of) {
    covidData::load_data(
      as_of = as_of,
      spatial_resolution = c("state", "national"),
      temporal_resolution = temporal_resolution,
      measure = measure
    ) %>%
      dplyr::mutate(
        as_of = as_of
      )
  }
)

output <- list()
for (location in all_locations) {
  location_name <- covidData::fips_codes %>%
    dplyr::filter(location == UQ(location)) %>%
    dplyr::pull(location_name)
  abbreviation <- covidData::fips_codes %>%
    dplyr::filter(location == UQ(location)) %>%
    dplyr::pull(abbreviation)
  
  for (as_of in as.character(all_as_ofs)) {
#    message(paste0(location_name, ", ", as_of))
    data_to_plot <- data_all_locations %>%
      dplyr::filter(
        location == UQ(location),
        as_of == UQ(as_of),
        date >= plot_start_date)

        # subset negative points
        data_to_plot_neg <- data_to_plot %>%
        filter(inc < 0)
      # outliers to plot
      outlier_to_plot <- death_outliers %>%
      dplyr::filter(
        location == UQ(location),
        date >= plot_start_date)
              
    if (nrow(data_to_plot) > 0) {
      p <- ggplot() +
        geom_line(data = data_to_plot,
          mapping = aes(x = date, y = inc)) +
        geom_point(data = data_to_plot, aes(date, inc))+
        geom_point(data = data_to_plot_neg, aes(date, inc), color = 'red')+
          geom_point(data = outlier_to_plot, aes(date, reported_inc), color = 'orange')+
        scale_x_date(
          breaks = data_to_plot %>%
            dplyr::filter(weekdays(date) == "Saturday") %>%
            dplyr::pull(date) %>%
            unique(),
          minor_breaks = NULL,
          limits = c(plot_start_date, plot_end_date)) +
        ggtitle(paste0(location_name," (", abbreviation,") as of ", as_of)) +
        theme_bw() +
        theme(
          axis.text.x = element_text(angle = 90, vjust = 0.5),
          panel.grid.major.x = element_line(colour = "darkgrey")
        )
      # plotly::ggplotly(p) Need help
      output[[length(output) + 1L]] <- '#### \n'

      pp <- plotly::ggplotly(p)
      output[[length(output) + 1L]] <- pp
      output[[length(output) + 1L]] <- cat('\n \n')
    }
  }
}

# Render the outputs
for(j in 1:length(output)){
  x <- output[[j]]

  if(inherits(x, "character")){
    cat("\n")
    cat(x)
  } else if(inherits(x, "knitr_kable")){
    cat("\n")
    print(x)
  }
  else {
    # print the html piece of the htmlwidgets
    cat("\n")
    cat(htmltools::renderTags(as_widget(x))$html)
  }
}
```

```{r}
# Attach the Dependencies since they do not get included with renderTags(...)$html
deps <- lapply(
  Filter(f = function(x){inherits(x,"htmlwidget")}, x = output),
  function(hw){
    htmltools::renderTags(hw)$dependencies
  }
)
htmltools::attachDependencies(x = htmltools::tagList(), value = unlist(deps,recursive=FALSE))
```