API tools faq
paste
Advertisement
MetricT

TN COVID-19 script, Apr 21

MetricT
Apr 21st, 2020
2,378
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
R 35.87 KB | None | 0 0
raw download clone embed print report
  1. ################################################################################
  2. ### Render a map of Tennessee using TN Dept. of Health data on COVID-19
  3. ### infections showing total cases, new cases, cases by age, and a map of
  4. ### confirmed infected.
  5. ###
  6. ### TN Dept. of Health COVID-19 website:
  7. ###     https://www.tn.gov/health/cedep/ncov.html
  8. ###
  9. ### Note: You will need a Census API key to access their data.  You
  10. ###       can get a key by going to this website:
  11. ###
  12. ###       https://api.census.gov/data/key_signup.html
  13. ###
  14. ### Once you have the key, just add it below.
  15. ################################################################################
  16.  
  17. ### Set the Census API key
  18. census_key <- "PUT_YOUR_CENSUS_API_KEY_HERE"
  19.  
  20. ### Let's start by cleaning the environment, it currently causes the graphs
  21. ### some problems if the environment is already populated
  22. rm(list = ls())
  23.  
  24. # Need to install the github version of googlesheets4
  25. # install.packages("devtools")
  26. devtools::install_github("tidyverse/googlesheets4")
  27.  
  28. ### Load the necessary libraries
  29. packages <- c("tidyverse", "dplyr", "readxl", "ggplot2", "sf", "rgeos", "TTR",
  30.               "scales", "cowplot", "viridis", "gridExtra", "tidycensus",
  31.               "googlesheets4")
  32.  
  33. new_packages <- packages[!(packages %in% installed.packages()[, "Package"])]
  34. if (length(new_packages)) install.packages(new_packages, quiet = TRUE)
  35. invisible(lapply(packages, "library", quietly = TRUE,
  36.                  character.only = TRUE, warn.conflicts = FALSE))
  37.  
  38. ### Set the Census API key
  39. census_api_key(census_key)
  40.  
  41. ### Enable map caching
  42. options(tigris_use_cache = TRUE)
  43.  
  44.  
  45. ################################################################################
  46. ### Read data from my spreadsheet
  47. ################################################################################
  48. ### Load COVID data from Google Sheets
  49. gs4_deauth()
  50. covid_data <- "https://docs.google.com/spreadsheets/d/1Opqs3DIYTlPbPhqrUa66RfKBBDVo0u5sntFM6uYZDzk/edit?usp=sharing"
  51. infected_df  <- read_sheet(covid_data, sheet = "Infected")
  52. deaths_df    <- read_sheet(covid_data, sheet = "Deaths")
  53. hospital_df  <- read_sheet(covid_data, sheet = "Hospitalizations")
  54. testres_df   <- read_sheet(covid_data, sheet = "Test_Results")
  55. testrates_df <- read_sheet(covid_data, sheet = "Testing_Rates") %>%
  56.                 mutate(County = gsub(" County", "", County))
  57. recover_df   <- read_sheet(covid_data, sheet = "Recovered")
  58. age_df       <- read_sheet(covid_data, sheet = "Age_Data")
  59.  
  60.  
  61. ################################################################################
  62. ### Subset the the data here (if desired)
  63. ################################################################################
  64. ### For the map, we only want the latest total for each county.
  65. non_county_fields <- c("Date", "Total", "Other_State_Country",
  66.                        "Unknown", "ERROR!!!")
  67.  
  68. ### All counties
  69. all_counties <-
  70.   colnames(infected_df) %>%
  71.   enframe() %>%
  72.   filter(!colnames(infected_df) %in% non_county_fields) %>%
  73.   select("value") %>%
  74.   deframe()
  75.  
  76. ### Counties in the Nashville Metropolitan Statistical Area
  77. nashville_msa <-
  78.   c("Cannon", "Cheatham", "Davidson", "Dickson", "Macon", "Maury", "Robertson",
  79.     "Rutherford", "Smith", "Sumner", "Trousdale", "Williamson", "Wilson")
  80.  
  81. middle_tn <- c(
  82.   "Bedford",    "Cannon",  "Cheatham",  "Clay",       "Coffee",    "Davidson",
  83.   "DeKalb",     "Dickson", "Fentress",  "Franklin",   "Giles",     "Grundy",
  84.   "Hickman",    "Houston", "Humphreys", "Jackson",    "Lawrence",  "Lewis",
  85.   "Lincoln",    "Macon",   "Marshall",  "Maury",      "Montgomery", "Moore",
  86.   "Overton",    "Perry",   "Pickett",   "Putnam",     "Robertson", "Rutherford",
  87.   "Sequatchie", "Smith",   "Stewart",   "Sumner",     "Trousdale", "Van Buren",
  88.   "Warren",     "Wayne",   "White",     "Williamson", "Wilson")
  89.  
  90.  
  91. ################################################################################
  92. ### Select the map you want
  93. ################################################################################
  94. #graph_counties <- nashville_msa
  95. #map_counties <- nashville_msa
  96. #location <- "Nashville MSA"
  97.  
  98. #graph_counties <- c("Davidson")
  99. #map_counties <- c("Davidson")
  100. #location <- "Davidson County"
  101.  
  102. #graph_counties <- middle_tn
  103. #map_counties <- middle_tn
  104. #location <- "Middle TN"
  105.  
  106. graph_counties <- c("Total")
  107. map_counties <- all_counties
  108. location <- "Tennessee"
  109.  
  110. ### Use TidyCensus to pull population data as well as map geometry
  111. ### from the Census bureau
  112. county_acs <-
  113.   get_acs(geography   = "county",
  114.           variables   = c(POP2018 = "B01003_001"),
  115.           state       = c("47"),
  116.           county      = map_counties,
  117.           year        = 2018,
  118.           geometry    = TRUE,
  119.           cache_table = TRUE) %>%
  120.   rename(POP2018 = estimate)
  121.  
  122. ### Split the Census data apart into a map...
  123. tn_map_df <-
  124.   county_acs %>%
  125.   select(GEOID, geometry)
  126.  
  127. ### And a population dataframe
  128. tn_pop_df <-
  129.   county_acs %>%
  130.   st_set_geometry(NULL) %>%
  131.   mutate(NAME = str_replace(NAME, " County, Tennessee", "")) %>%
  132.   select(GEOID, NAME, POP2018) %>%
  133.   rename(County = NAME)
  134.  
  135.  
  136. ################################################################################
  137. ### A few visual configuration settings
  138. ################################################################################
  139. map_palette <- c("#c4c4c4", viridis(n = 1000, direction = -1))
  140. map_fontsize <- 3
  141. graph_color <- "darkseagreen4"
  142. geom_thickness <- 0.25
  143. geom_age_orientation <- coord_flip()
  144. ENABLE_LOG_SCALE <- TRUE
  145.  
  146. ### If TRUE, graph total cases, new cases, and total deaths on a log scale
  147. ### If FALSE, graph them normally
  148. if (isTRUE(ENABLE_LOG_SCALE)) {
  149.      graph_log10_opts1 <- scale_y_log10(breaks = c(1, 10, 100, 1000))
  150.      graph_log10_opts2 <- annotation_logticks()
  151. } else {
  152.      graph_log10_opts1 <- geom_blank()
  153.      graph_log10_opts2 <- geom_blank()
  154. }
  155.  
  156.  
  157. ################################################################################
  158. ### Adjusting the geometry of our map
  159. ################################################################################
  160. ### Massage the state/county map data a bit
  161. state_mapdata <- tn_pop_df
  162.  
  163. ### Make a copy of the state geometry
  164. counties <- state_mapdata
  165.  
  166. ### Find the center of each county so we can add the number of infected
  167. county_centers <-
  168.   tn_map_df$geometry %>%
  169.   as_Spatial() %>%
  170.   gCentroid(byid = TRUE)
  171.  
  172. ### The centroid function is very good, but occasionally can give less-than-
  173. ### perfect results, often due to counties with extreme concave shape.   We
  174. ### can use "nudges" to adjust the location of the number to make it look good.
  175. ### Default nudge is 0.
  176. counties$nudge_y <- 0
  177. counties$nudge_x <- 0
  178.  
  179. counties$nudge_y[counties$County == "Chester"]   <-  0.035
  180. counties$nudge_y[counties$County == "Houston"]   <-  0.015
  181. counties$nudge_y[counties$County == "Putnam"]    <-  0.025
  182. counties$nudge_y[counties$County == "Loudon"]    <-  0.015
  183. counties$nudge_y[counties$County == "Anderson"]  <- -0.015
  184. counties$nudge_x[counties$County == "Pickett"]   <- -0.065
  185. counties$nudge_x[counties$County == "Unicoi"]    <- -0.070
  186. counties$nudge_y[counties$County == "Unicoi"]    <- -0.035
  187. counties$nudge_x[counties$County == "Hancock"]   <- -0.035
  188. counties$nudge_x[counties$County == "Trousdale"] <- -0.020
  189. counties$nudge_x[counties$County == "Lake"]      <-  0.030
  190. counties$nudge_x[counties$County == "Moore"]     <-  0.005
  191. counties$nudge_y[counties$County == "Moore"]     <-  0.015
  192.  
  193.  
  194. ################################################################################
  195. ### Math section for maps
  196. ################################################################################
  197. ### Get the number of total confirmed cases by county
  198. totalinfected_by_county <-
  199.   infected_df %>%
  200.   arrange(Date) %>%
  201.   tail(n = 1) %>%
  202.   select(-all_of(non_county_fields)) %>%
  203.   gather(Date) %>%
  204.   rename(County = Date) %>%
  205.   rename(totalinfected = value)
  206.  
  207. ### Get the number of new confirmed cases by county
  208. new_date <-
  209.   infected_df %>%
  210.   pull(Date) %>%
  211.   max() %>%
  212.   as_tibble() %>%
  213.   rename(Date = value) %>%
  214.   pull()
  215.  
  216. newinfected_by_county <-
  217.   infected_df %>%
  218.   arrange(Date) %>%
  219.   tail(n = 2) %>%
  220.   select(-all_of(non_county_fields))
  221.  
  222. newinfected_by_county <-
  223.   (newinfected_by_county[2, ] - newinfected_by_county[1, ]) %>%
  224.   as_tibble() %>%
  225.   add_column(new_date, .before = 1) %>%
  226.   rename(Date = new_date) %>%
  227.   gather(Date) %>%
  228.   rename(County = Date) %>%
  229.   rename(newinfected = value)
  230.  
  231. ### Get the number of total deaths by county
  232. totaldeaths_by_county <-
  233.   deaths_df %>%
  234.   filter(Date > "2020-03-31") %>%
  235.   head(n = 1) %>%
  236.   select("Date", all_of(all_counties)) %>%
  237.   gather(Date) %>%
  238.   rename(County = Date) %>%
  239.   rename(totaldeaths = value)
  240.  
  241. ### Get the number of new deaths by county
  242. new_date <-
  243.   deaths_df %>%
  244.   select("Date") %>%
  245.   filter(!is.na(Date)) %>%
  246.   pull(Date) %>%
  247.   max() %>%
  248.   as_tibble() %>%
  249.   pull()
  250.  
  251. newdeaths_by_county <-
  252.   deaths_df %>%
  253.   filter(!is.na(Date)) %>%
  254.   arrange(Date) %>%
  255.   tail(n = 2) %>%
  256.   select(-all_of(non_county_fields))
  257.  
  258. newdeaths_by_county <-
  259.   (newdeaths_by_county[2, ] - newdeaths_by_county[1, ]) %>%
  260.   as_tibble() %>%
  261.   add_column(new_date, .before = 1) %>%
  262.   rename(Date = new_date) %>%
  263.   gather(Date) %>%
  264.   rename(County = Date) %>%
  265.   rename(newdeaths = value)
  266.  
  267. ### Get the number of total deaths by county
  268. recovered_by_county <-
  269.   recover_df %>%
  270.   filter(Date > "2020-03-31") %>%
  271.   arrange(Date) %>%
  272.   tail(n = 1) %>%
  273.   select("Date", all_of(all_counties)) %>%
  274.   gather(Date) %>%
  275.   rename(County = Date) %>%
  276.   rename(recovered = value)
  277.  
  278. ### Still sick
  279. sick_by_county <-
  280.   totalinfected_by_county %>%
  281.   left_join(recovered_by_county, by = "County") %>%
  282.   left_join(totaldeaths_by_county, by = "County") %>%
  283.   mutate(sick = totalinfected - recovered - totaldeaths) %>%
  284.   select(County, sick)
  285.  
  286. ### Combine the data
  287. stats_df <-
  288.   totalinfected_by_county %>%
  289.   left_join(newinfected_by_county, by = "County") %>%
  290.   left_join(totaldeaths_by_county, by = "County") %>%
  291.   left_join(newdeaths_by_county, by = "County")   %>%
  292.   left_join(sick_by_county, by = "County") %>%
  293.   left_join(recovered_by_county, by = "County")
  294.  
  295. ### Combine all the above together
  296. counties <-
  297.   counties %>%
  298.   left_join(tn_map_df, by = "GEOID") %>%
  299.   left_join(stats_df,  by = "County") %>%
  300.   mutate(newinfected = if_else(newinfected == -1, 0, newinfected))
  301.  
  302.  
  303. ################################################################################
  304. ### Map of total COVID-19 cases
  305. ################################################################################
  306. ### We want to add text with the number of infected in each county to that
  307. ### county on the map.   But we *don't* want to label counties that don't yet
  308. ### have any confirmed cases.
  309. totalinfected_label <- counties$totalinfected
  310. totalinfected_label[totalinfected_label == 0] <- ""
  311.  
  312. ### By default use black for the font, but if the value is over 1/2's of the way
  313. ### to the max, use the white font instead as it looks nicer
  314. frac <- 0.5 * max(counties$totalinfected)
  315. counties$textcolor <- if_else(counties$totalinfected >= frac, "white", "black")
  316.  
  317. ### Map:  COVID-19 confirmed infected per county
  318. p_map_totalinfected <-
  319.   ggplot(counties) +
  320.   theme_void() +
  321.   theme(legend.title = element_blank()) +
  322.   theme(plot.title = element_text(hjust = 0.5)) +
  323.   theme(legend.position = "none") +
  324.   geom_sf(data = counties$geometry, aes(fill = counties$totalinfected),
  325.           size = geom_thickness) +
  326.   geom_text(data = counties, size = map_fontsize, fontface = "bold",
  327.             color = counties$textcolor,
  328.             aes(x     = county_centers$x,
  329.                 y     = county_centers$y,
  330.                 label = totalinfected_label),
  331.             nudge_x = counties$nudge_x,
  332.             nudge_y = counties$nudge_y) +
  333.   labs(title = "Total Cases") +
  334.   scale_fill_gradientn(colours = map_palette)
  335.  
  336.  
  337. ################################################################################
  338. ### Map of new COVID-19 cases
  339. ################################################################################
  340. ### We want to add text with the number of infected in each county to that
  341. ### county on the map.   But we *don't* want to label counties that don't yet
  342. ### have any confirmed cases.
  343. newinfected_label <- counties$newinfected
  344. newinfected_label[newinfected_label <= 0] <- ""
  345.  
  346. ### By default use black for the font, but if the value is over 1/2's of the way
  347. ### to the max, use the white font instead as it looks nicer
  348. frac <- 0.5 * max(counties$newinfected)
  349. counties$textcolor <- if_else(counties$newinfected >= frac, "white", "black")
  350.  
  351. ### Map:  COVID-19 confirmed new infected per county
  352. p_map_newinfected <-
  353.   ggplot(counties) +
  354.   theme_void() +
  355.   theme(legend.title = element_blank()) +
  356.   theme(plot.title = element_text(hjust = 0.5)) +
  357.   theme(legend.position = "none") +
  358.   geom_sf(data = counties$geometry, aes(fill = counties$newinfected),
  359.           size = geom_thickness) +
  360.   geom_text(data = counties, size = map_fontsize, fontface = "bold",
  361.             color = counties$textcolor,
  362.             aes(x     = county_centers$x,
  363.                 y     = county_centers$y,
  364.                 label = newinfected_label),
  365.             nudge_x = counties$nudge_x,
  366.             nudge_y = counties$nudge_y) +
  367.   labs(title = "New Cases") +
  368.   scale_fill_gradientn(colours = map_palette)
  369.  
  370.  
  371. ################################################################################
  372. ### Map of deaths by county
  373. ################################################################################
  374. total_deaths_label <- counties$totaldeaths
  375. total_deaths_label[total_deaths_label == 0] <- ""
  376.  
  377. ### By default use black for the font, but if the value is over 1/2's of the way
  378. ### to the max, use the white font instead as it looks nicer
  379. frac <- 0.5 * max(counties$totaldeaths)
  380. counties$textcolor <- if_else(counties$totaldeaths >= frac, "white", "black")
  381.  
  382. p_map_totaldeaths <- ggplot(counties) +
  383.   theme_void() +
  384.   theme(legend.title = element_blank()) +
  385.   theme(plot.title = element_text(hjust = 0.5)) +
  386.   theme(legend.position = "none") +
  387.   geom_sf(data = counties$geometry, size = geom_thickness,
  388.           aes(fill = counties$totaldeaths)) +
  389.   geom_text(data = counties, size = map_fontsize, fontface = "bold",
  390.             color = counties$textcolor,
  391.             aes(x     = county_centers$x,
  392.                 y     = county_centers$y,
  393.                 label = total_deaths_label),
  394.             nudge_x = counties$nudge_x,
  395.             nudge_y = counties$nudge_y) +
  396.   labs(title = "Total Deaths") +
  397.   scale_fill_gradientn(colours = map_palette)
  398.  
  399. new_deaths_label <- counties$newdeaths
  400. new_deaths_label[new_deaths_label == 0] <- ""
  401.  
  402. ### By default use black for the font, but if the value is over 1/2's of the way
  403. ### to the max, use the white font instead as it looks nicer
  404. frac <- 0.5 * max(counties$newdeaths)
  405. counties$textcolor <- if_else(counties$newdeaths > frac, "white", "black")
  406.  
  407. p_map_newdeaths <- ggplot(counties) +
  408.   theme_void() +
  409.   theme(legend.title = element_blank()) +
  410.   theme(plot.title = element_text(hjust = 0.5)) +
  411.   theme(legend.position = "none") +
  412.   geom_sf(data = counties$geometry, aes(fill = counties$newdeaths),
  413.           size = geom_thickness) +
  414.   geom_text(data = counties, size = map_fontsize, fontface = "bold",
  415.             color = counties$textcolor,
  416.             aes(x     = county_centers$x,
  417.                 y     = county_centers$y,
  418.                 label = new_deaths_label),
  419.             nudge_x = counties$nudge_x,
  420.             nudge_y = counties$nudge_y) +
  421.   labs(title = "New Deaths") +
  422.   scale_fill_gradientn(colours = map_palette)
  423.  
  424.  
  425. ################################################################################
  426. ### Map of deaths per 100k by county
  427. ################################################################################
  428. counties$deaths_per100k <- round(100000 * counties$totaldeaths /
  429.                                    counties$POP2018, 3)
  430. deaths100k_label <- round(counties$deaths_per100k, 1)
  431. deaths100k_label[deaths100k_label == 0] <- ""
  432.  
  433. ### By default use black for the font, but if the value is over 1/2's of the way
  434. ### to the max, use the white font instead as it looks nicer
  435. frac <- 0.5 * max(counties$deaths_per100k)
  436. counties$textcolor <- if_else(counties$deaths_per100k >= frac, "white", "black")
  437.  
  438. p_map_totaldeaths_100k <- ggplot(counties) +
  439.   theme_void() +
  440.   theme(legend.title = element_blank()) +
  441.   theme(plot.title = element_text(hjust = 0.5)) +
  442.   theme(legend.position = "none") +
  443.   geom_sf(data = counties$geometry, aes(fill = counties$deaths_per100k),
  444.           size = geom_thickness) +
  445.   geom_text(data = counties, size = map_fontsize, fontface = "bold",
  446.             color = counties$textcolor,
  447.             aes(x     = county_centers$x,
  448.                 y     = county_centers$y,
  449.                 label = deaths100k_label),
  450.             nudge_x = counties$nudge_x,
  451.             nudge_y = counties$nudge_y) +
  452.   labs(title = "Total Deaths per 100k") +
  453.   scale_fill_gradientn(colours = map_palette)
  454.  
  455.  
  456. ################################################################################
  457. ### Map of total COVID-19 cases per 100k
  458. ################################################################################
  459. counties$infected_percapita <-
  460.   round(100000 * counties$totalinfected / counties$POP2018, 0)
  461.  
  462. frac <- 0.5 * max(counties$infected_percapita)
  463. counties$textcolor <- if_else(counties$infected_percapita >= frac,
  464.                               "white", "black")
  465.  
  466. infected_label_per  <- round(counties$infected_percapita, 1)
  467. infected_label_per[infected_label_per == 0] <- ""
  468.  
  469. p_map_totalinfected_per <- ggplot(counties) +
  470.    theme_void() +
  471.    theme(plot.title = element_text(hjust = 0.5)) +
  472.    theme(plot.caption = element_text(hjust = 0.5)) +
  473.    theme(legend.title = element_blank()) +
  474.    theme(legend.position = "none") +
  475.    geom_sf(data = counties$geometry, size = geom_thickness,
  476.              aes(fill = counties$infected_percapita)) +
  477.    geom_text(data = counties, size = map_fontsize, fontface = "bold",
  478.              color = counties$textcolor,
  479.              aes(x     = county_centers$x,
  480.                  y     = county_centers$y,
  481.                  label = infected_label_per),
  482.                nudge_x = counties$nudge_x,
  483.                nudge_y = counties$nudge_y) +
  484.    labs(title = "Total Cases per 100k") +
  485.      scale_fill_gradientn(colours = map_palette)
  486.  
  487.  
  488. ################################################################################
  489. ### Map of people still sick (not recovered) by county
  490. ################################################################################
  491. frac <- 0.5 * max(counties$sick)
  492. counties$textcolor <- if_else(counties$sick >= frac, "white", "black")
  493.  
  494. sick_label_per  <- counties$sick
  495. sick_label_per[sick_label_per == 0] <- ""
  496.  
  497. p_map_sick <- ggplot(counties) +
  498.   theme_void() +
  499.   theme(plot.title = element_text(hjust = 0.5)) +
  500.   theme(plot.caption = element_text(hjust = 0.5)) +
  501.   theme(legend.title = element_blank()) +
  502.   theme(legend.position = "none") +
  503.   geom_sf(data = counties$geometry, size = geom_thickness,
  504.           aes(fill = counties$sick)) +
  505.   geom_text(data = counties, size = map_fontsize, fontface = "bold",
  506.             color = counties$textcolor,
  507.             aes(x     = county_centers$x,
  508.                 y     = county_centers$y,
  509.                 label = sick_label_per),
  510.             nudge_x = counties$nudge_x,
  511.             nudge_y = counties$nudge_y) +
  512.   labs(title = "Still Sick") +
  513.   scale_fill_gradientn(colours = map_palette)
  514.  
  515.  
  516. ################################################################################
  517. ### Map of people recovered by county
  518. ################################################################################
  519. frac <- 0.5 * max(counties$recovered)
  520. counties$textcolor <- if_else(counties$recovered >= frac, "white", "black")
  521.  
  522. recovered_label_per  <- counties$recovered
  523. recovered_label_per[recovered_label_per == 0] <- ""
  524.  
  525. p_map_recovered <- ggplot(counties) +
  526.   theme_void() +
  527.   theme(plot.title = element_text(hjust = 0.5)) +
  528.   theme(plot.caption = element_text(hjust = 0.5)) +
  529.   theme(legend.title = element_blank()) +
  530.   theme(legend.position = "none") +
  531.   geom_sf(data = counties$geometry, size = geom_thickness,
  532.           aes(fill = counties$recovered)) +
  533.   geom_text(data = counties, size = map_fontsize, fontface = "bold",
  534.             color = counties$textcolor,
  535.             aes(x     = county_centers$x,
  536.                 y     = county_centers$y,
  537.                 label = recovered_label_per),
  538.             nudge_x = counties$nudge_x,
  539.             nudge_y = counties$nudge_y) +
  540.   labs(title = "Recovered") +
  541.   scale_fill_gradientn(colours = map_palette)
  542.  
  543.  
  544. ################################################################################
  545. ### Math section for charts
  546. ################################################################################
  547. infected_data <-
  548.   infected_df %>%
  549.   arrange(Date) %>%
  550.   select("Date", all_of(graph_counties)) %>%
  551.   mutate(Total = rowSums(select(., !starts_with("Date")))) %>%
  552.   select(Date, Total) %>%
  553.   rename(totalinfected = Total) %>%
  554.   rename(date = Date) %>%
  555.   filter(totalinfected != 0) %>%
  556.   mutate(newinfected = c(totalinfected[1], diff(totalinfected))) %>%
  557.   mutate(numdays = difftime(as.POSIXct(date),
  558.                             as.POSIXct(as.Date(min(date)) - 1, tz = "UTC"),
  559.                             units = "days")) %>%
  560.   mutate(numdays = as.numeric(numdays))
  561.  
  562. hospital_data <-
  563.   hospital_df %>%
  564.   arrange(Date) %>%
  565.   mutate(numdays = difftime(as.POSIXct(Date),
  566.                             as.POSIXct(as.Date(min(Date)) - 1, tz = "UTC"),
  567.                             units = "days")) %>%
  568.   mutate(numdays = as.numeric(numdays)) %>%
  569.   rename(total_hospitalizations = Hospitalizations) %>%
  570.   mutate(new_hospitalizations = c(total_hospitalizations[1],
  571.                                   diff(total_hospitalizations)))
  572.  
  573. recovered_data <-
  574.   recover_df %>%
  575.   rename(Total = Recovered) %>%
  576.   select("Date", all_of(graph_counties)) %>%
  577.   mutate(Total = rowSums(select(., !starts_with("Date")))) %>%
  578.   select(Date, Total) %>%
  579.   rename(totalrecovered = Total) %>%
  580.   rename(date = Date) %>%
  581.   filter(totalrecovered != 0) %>%
  582.   arrange(date) %>%
  583.   mutate(newrecovered = c(totalrecovered[1], diff(totalrecovered))) %>%
  584.   mutate(numdays = difftime(as.POSIXct(date),
  585.                             as.POSIXct(as.Date(min(date)) - 1, tz = "UTC"),
  586.                             units = "days")) %>%
  587.   mutate(numdays = as.numeric(numdays))
  588.  
  589. deaths_data <-
  590.   deaths_df %>%
  591.   na.omit() %>%
  592.   arrange(Date) %>%
  593.   select("Date", all_of(graph_counties)) %>%
  594.   mutate(Total = rowSums(select(., !starts_with("Date")))) %>%
  595.   mutate(New = c(1, diff(Total))) %>%
  596.   select(Date, Total, New) %>%
  597.   filter(Total != 0) %>%
  598.   rename(totaldeaths = Total) %>%
  599.   rename(newdeaths = New) %>%
  600.   rename(date = Date) %>%
  601.   mutate(numdays = difftime(as.POSIXct(date),
  602.                             as.POSIXct(as.Date(min(date)) - 1, tz = "UTC"),
  603.                             units = "days")) %>%
  604.   mutate(numdays = as.numeric(numdays))
  605.  
  606. sick_data <-
  607.   infected_data %>%
  608.   left_join(recovered_data, by = "date") %>%
  609.   left_join(deaths_data,   by = "date") %>%
  610.   select(date, totalinfected, totaldeaths, totalrecovered) %>%
  611.   filter(date >= as.Date("2020-04-10")) %>%
  612.   filter(!is.na(totaldeaths)) %>%
  613.   filter(!is.na(totalrecovered)) %>%
  614.   filter(!is.na(totalinfected)) %>%
  615.   mutate(current_sick = totalinfected - totaldeaths - totalrecovered) %>%
  616.   select(date, current_sick)
  617.  
  618.  
  619. ################################################################################
  620. ### Graph:  COVID-19 Total Confirmed Cases
  621. ################################################################################
  622. ### Use non-linear fitting to do the exponential fit
  623. fit_totalinfected <-
  624.   nls(totalinfected ~ a * exp(b * numdays),
  625.       data = infected_data,
  626.       start = list(a = 334,
  627.                    b = 0.0677))
  628.  
  629. ### Extract the regression coefficents so we can add the regression formula
  630. ### to the graph
  631. intercept  <- coef(fit_totalinfected)[1]
  632. d_constant <- coef(fit_totalinfected)[2]
  633.  
  634. ### Add the fit to our dataframe
  635. infected_data <-
  636.   infected_data %>%
  637.   mutate(fit_totalinfected = intercept *
  638.            exp(d_constant * infected_data$numdays))
  639.  
  640. ### Compute the doubling time
  641. doubling_time <- log(2) / d_constant
  642.  
  643. ### Compute the "order-of-magnitude" time
  644. magnitude_time <- log(10) / d_constant
  645.  
  646. ### Growth rate
  647. growth_rate <- exp(log(2) / doubling_time) - 1
  648.  
  649. ### What's the current number of total infected cases
  650. total_num <-
  651.   infected_data %>%
  652.   tail(n = 1) %>%
  653.   select("totalinfected") %>%
  654.   pull()
  655.  
  656. ### What's the date of the last data point
  657. data_date <- infected_data  %>% select("date") %>% tail(n = 1) %>% pull()
  658.  
  659. p_total <- ggplot(data = infected_data) +
  660.   theme_classic() +
  661.   theme(axis.text.x = element_text(angle = 45)) +
  662.   theme(axis.text.x = element_text(vjust = 0.7)) +
  663.   theme(axis.text.x = element_text(hjust = 0.8)) +
  664.   theme(legend.title = element_blank()) +
  665.   theme(legend.position = "none") +
  666.   geom_point(data = infected_data, shape = 19, size = 0.5,
  667.              aes(x = as.Date(date), y = totalinfected), color = "black") +
  668.   geom_line(data = infected_data, color = graph_color,
  669.             aes(x = as.Date(date), y = fit_totalinfected)) +
  670.   scale_x_date(date_breaks = "3 days", date_labels = "%m/%d") +
  671.   graph_log10_opts1 +
  672.   graph_log10_opts2 +
  673.   labs(title = paste("Total Cases: ", total_num, sep = ""), x = "", y = "")
  674.  
  675.  
  676. ################################################################################
  677. ### Graph:  COVID-19 New Confirmed Cases
  678. ################################################################################
  679. ### If we assume infected = a * B^(c * days), then
  680. ###
  681. ###  d(infected)/dt = c * log_e(B) * infected
  682. ###
  683. ### We can use the same constants from the total fit on the new fit.   If new
  684. ### cases are still growing exponentially, it will work.   And when it doesn't
  685. ### work, that's a sign that we're either "flattening the curve", or something
  686. ### else weird is going on
  687.  
  688. ### Add the fit to our dataframe
  689. infected_data <-
  690.   infected_data %>%
  691.   mutate(fit_newinfected = fit_totalinfected * d_constant)
  692.  
  693. ### What's the current number of total infected cases
  694. new_num <- infected_data %>% tail(n = 1) %>% select("newinfected") %>% pull()
  695. SMA_num <- infected_data$newinfected %>% SMA(n = 7) %>% tail(n = 1) %>% round()
  696.  
  697. ### Graph:  new COVID-19 Cases
  698. p_new <- ggplot(data = infected_data) +
  699.   theme_classic() +
  700.   theme(axis.text.x = element_text(angle = 45)) +
  701.   theme(axis.text.x = element_text(vjust = 0.7)) +
  702.   theme(axis.text.x = element_text(hjust = 0.8)) +
  703.   theme(legend.title = element_blank()) +
  704.   theme(legend.position = "none") +
  705.   geom_point(data = infected_data, shape = 19, size = 0.5,
  706.              aes(x = as.Date(date), y = newinfected), color = "black") +
  707.   geom_line(data = infected_data, color = graph_color,
  708.             size = 1.0, aes(x = as.Date(date) - 3,
  709.                             y = SMA(newinfected, n = 7))) +
  710.   scale_x_date(date_breaks = "3 days", date_labels = "%m/%d") +
  711.   graph_log10_opts1 +
  712.   graph_log10_opts2 +
  713.   labs(title = paste("New Cases: ", new_num, ", SMA: ", SMA_num, sep = ""),
  714.        x = "", y = "")
  715.  
  716.  
  717. ################################################################################
  718. ### Graph:  COVID-19 Cases [ log(Total) vs log(New) ]
  719. ### Reference:  https://aatishb.com/covidtrends/
  720. ################################################################################
  721. p_total_new <- ggplot(data = infected_data,
  722.                       aes(x = totalinfected, y = newinfected)) +
  723.   theme_classic() +
  724.   theme(axis.text.x = element_text(angle = 45)) +
  725.   theme(axis.text.x = element_text(vjust = 0.7)) +
  726.   theme(axis.text.x = element_text(hjust = 0.8)) +
  727.   theme(legend.position = "none") +
  728.   geom_point(data = infected_data, shape = 19, size = 0.5,
  729.              aes(x = totalinfected, y = newinfected), color = "black") +
  730.   geom_smooth(method = "loess", formula = y ~ x,
  731.               se = FALSE, size = 0.5, color = graph_color) +
  732.   scale_x_log10(breaks = c(1, 10, 100, 1000)) +
  733.   scale_y_log10(breaks = c(1, 10, 100, 1000)) +
  734.   annotation_logticks() +
  735.   labs(title = "Log(New vs Total)", x = "", y = "")
  736.  
  737.  
  738. ################################################################################
  739. ### Graph:  COVID-19 Hospitalizations
  740. ################################################################################
  741. hospital_num <-
  742.   hospital_data %>%
  743.   arrange(Date) %>%
  744.   tail(n = 1) %>%
  745.   select("total_hospitalizations") %>%
  746.   pull()
  747.  
  748. p_total_hospital <- ggplot(data = hospital_data,
  749.                            aes(x = as.Date(Date), y = total_hospitalizations)) +
  750.   theme_classic() +
  751.   theme(axis.text.x = element_text(angle = 45)) +
  752.   theme(axis.text.x = element_text(vjust = 0.7)) +
  753.   theme(axis.text.x = element_text(hjust = 0.8)) +
  754.   theme(legend.title = element_blank()) +
  755.   theme(legend.position = "none") +
  756.   geom_bar(stat = "identity", fill = graph_color) +
  757.   scale_x_date(date_breaks = "3 days", date_labels = "%m/%d") +
  758.   labs(title = paste("TN Hospitalizations: ", hospital_num, sep = ""),
  759.        x = "", y = "")
  760.  
  761.  
  762. ################################################################################
  763. ### Graph:  COVID-19 Total Recovered
  764. ################################################################################\
  765. new_num <-
  766.   recovered_data %>%
  767.   tail(n = 1) %>%
  768.   select("totalrecovered") %>%
  769.   pull()
  770.  
  771. p_recover <- ggplot(data = recovered_data,
  772.                     aes(x = as.Date(date), y = totalrecovered)) +
  773.   theme_classic() +
  774.   theme(axis.text.x = element_text(angle = 45)) +
  775.   theme(axis.text.x = element_text(vjust = 0.7)) +
  776.   theme(axis.text.x = element_text(hjust = 0.8)) +
  777.   theme(legend.title = element_blank()) +
  778.   theme(legend.position = "none") +
  779.   geom_point(data = recovered_data, shape = 19, size = 0.5,
  780.              aes(x = as.Date(date), y = totalrecovered), color = "black") +
  781.   scale_x_date(date_breaks = "3 days", date_labels = "%m/%d") +
  782.   labs(title = paste("Total Recovered: ", new_num, sep = ""), x = "", y = "")
  783.  
  784.  
  785. ################################################################################
  786. ### Graph:  COVID-19 Total Infected by Age
  787. ################################################################################
  788. p_age_infected <- ggplot(data = age_df, aes(x = infected_statewide, y = age)) +
  789.   theme_classic() +
  790.   theme(axis.text.x = element_text(angle = 45)) +
  791.   theme(axis.text.x = element_text(vjust = 0.7)) +
  792.   theme(axis.text.x = element_text(hjust = 0.8)) +
  793.   geom_age_orientation +
  794.   geom_bar(stat = "identity", fill = graph_color) +
  795.   labs(title = "TN Infected by Age", x = "", y = "")
  796.  
  797.  
  798. ################################################################################
  799. ### Graph:  COVID-19 Total Infected by Age
  800. ################################################################################
  801. p_age_deaths <- ggplot(data = age_df, aes(x = deaths_statewide, y = age)) +
  802.   theme_classic() +
  803.   theme(axis.text.x = element_text(angle = 45)) +
  804.   theme(axis.text.x = element_text(vjust = 0.7)) +
  805.   theme(axis.text.x = element_text(hjust = 0.8)) +
  806.   geom_age_orientation +
  807.   geom_bar(stat = "identity", fill = graph_color) +
  808.   labs(title = "TN Deaths by Age", x = "", y = "")
  809.  
  810.  
  811. ################################################################################
  812. ### Graph:  COVID-19 Total/New Deaths
  813. ################################################################################
  814. ### Use non-linear fitting to do the exponential fit
  815. fit_totaldeaths <- nls(totaldeaths ~ a * exp(b * numdays),
  816.                   data = deaths_data, start = list(a = 12.51, b = 0.082))
  817.  
  818. ### Extract the regression coefficents so we can add the regression formula
  819. ### to the graph
  820. intercept  <- coef(fit_totaldeaths)[1]
  821. d_constant <- coef(fit_totaldeaths)[2]
  822.  
  823. ### Add fit for total/new deaths
  824. deaths_data <-
  825.   deaths_data %>%
  826.   mutate(fit_totaldeaths = intercept * exp(d_constant * numdays)) %>%
  827.   mutate(fit_newdeaths   = fit_totaldeaths * d_constant)
  828.  
  829. p_totaldeaths <- ggplot(data = deaths_data) +
  830.   theme_classic() +
  831.   theme(axis.text.x = element_text(angle = 45)) +
  832.   theme(axis.text.x = element_text(vjust = 0.7)) +
  833.   theme(axis.text.x = element_text(hjust = 0.8)) +
  834.   theme(legend.title = element_blank()) +
  835.   theme(legend.position = "none") +
  836.   geom_point(shape = 19, size = 0.5,
  837.              aes(x = as.Date(date), y = totaldeaths), color = "black") +
  838.   geom_line(color = graph_color,
  839.             aes(x = as.Date(date), y = fit_totaldeaths)) +
  840.   scale_x_date(date_breaks = "3 days", date_labels = "%m/%d") +
  841.   graph_log10_opts1 +
  842.   graph_log10_opts2 +
  843.   labs(title = paste("Total Deaths: ", max(deaths_data$totaldeaths),  sep = ""),
  844.        x = "", y = "")
  845.  
  846. p_newdeaths <- ggplot(data = deaths_data) +
  847.   theme_classic() +
  848.   theme(axis.text.x = element_text(angle = 45)) +
  849.   theme(axis.text.x = element_text(vjust = 0.7)) +
  850.   theme(axis.text.x = element_text(hjust = 0.8)) +
  851.   theme(legend.title = element_blank()) +
  852.   theme(legend.position = "none") +
  853.   geom_point(shape = 19, size = 0.5,
  854.              aes(x = as.Date(date), y = newdeaths), color = "black") +
  855.   geom_line(color = graph_color,
  856.             aes(x = as.Date(date), y = fit_newdeaths)) +
  857.   scale_x_date(date_breaks = "3 days", date_labels = "%m/%d") +
  858.   graph_log10_opts1 +
  859.   graph_log10_opts2 +
  860.   labs(title = paste("New Deaths: ", max(deaths_data$newdeaths),  sep = ""),
  861.        x = "", y = "")
  862.  
  863.  
  864. ################################################################################
  865. ### Graph:  COVID-19 Current Sick
  866. ################################################################################
  867. new_num <- sick_data %>% tail(n = 1) %>% select("current_sick") %>% pull()
  868.  
  869. p_sick <- ggplot(data = sick_data, aes(x = as.Date(date), y = current_sick)) +
  870.   theme_classic() +
  871.   theme(axis.text.x = element_text(angle = 45)) +
  872.   theme(axis.text.x = element_text(vjust = 0.7)) +
  873.   theme(axis.text.x = element_text(hjust = 0.8)) +
  874.   theme(legend.title = element_blank()) +
  875.   theme(legend.position = "none") +
  876.   geom_point(data = sick_data, shape = 19, size = 0.5,
  877.              aes(x = as.Date(date), y = current_sick), color = "black") +
  878.   scale_x_date(date_breaks = "3 days", date_labels = "%m/%d") +
  879.   labs(title = paste("Currently Sick: ", new_num, sep = ""), x = "", y = "")
  880.  
  881.  
  882. ################################################################################
  883. ### Arrange the graphs into the final product
  884. ################################################################################
  885. # What's the date of the last data point
  886. data_date <- infected_df %>% arrange(Date) %>% tail(n = 1) %>% pull("Date")
  887.  
  888. # Title
  889. title_string <- paste("COVID-19 Confirmed Cases in ", location,
  890.                       " [", as.Date(data_date), "]", sep = "")
  891.  
  892. # Footer
  893. footer_string <- paste("Source:  Tennessee Department of Health\n",
  894.                        "https://www.tn.gov/health/cedep/ncov.html", sep = "")
  895.  
  896. grob_charts <-
  897.   plot_grid(p_total,        p_new,        p_totaldeaths,    p_recover,   p_sick,
  898.             p_age_infected, p_age_deaths, p_total_hospital, p_total_new, NA,
  899.             ncol = 5, nrow = 2, align = "hv")
  900.  
  901. grob_maps <-
  902.   plot_grid(p_map_recovered,    p_map_totalinfected,
  903.             p_map_sick,         p_map_totalinfected_per,
  904.             p_map_totaldeaths,  p_map_newinfected,
  905.             ncol = 2, nrow = 3, align = "hv")
  906.  
  907. charts <- plot_grid(grob_maps,
  908.                     grob_charts,
  909.                     ncol = 1, nrow = 2,
  910.                     rel_widths  = c(1.2, 1),
  911.                     rel_heights = c(1, 1))
  912.  
  913. title <- ggdraw() + draw_label(title_string, fontface = "bold")
  914. Sys.sleep(2)
  915.  
  916. footer <- ggdraw() + draw_label(footer_string, size = 10)
  917. Sys.sleep(2)
  918.  
  919. final <- plot_grid(title, charts, footer, nrow = 3,
  920.                    rel_heights = c(0.05, 1, 0.05))
  921. Sys.sleep(2)
  922.  
  923. print(final)
  924.  
  925. ### Save image to file and we're done!
  926. scale <- 1.3
  927. width <- round(1300 * scale)
  928. height <- round(900 * scale)
  929. png("TN_COVID_Viridis.png", width = width, height = height)
  930. plot(final)
  931. dev.off()
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement
Public Pastes
Advertisement
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!