covid-19 cases growth rates

corona-growth.R

needs(dplyr, readr, reshape2, tidyr, zoo)

read_csv("https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv") %>%
  rename(country = `Country/Region`,
         province = `Province/State`) %>%
  select(-Lat, -Long) %>%
  pivot_longer(-c(country, province), names_to = "date", values_to = "confirmed") %>%
  mutate(date = as.Date(date, "%m/%d/%y")) %>%
  mutate(country = ifelse(country == "China" & province == "Hong Kong",
                          "Hong Kong", country)) %>%
  select(-province) %>%
  arrange(date) %>%
  group_by(date, country) %>%
  mutate(confirmed = sum(confirmed)) %>%
  distinct() %>%
  ungroup() %>%
  filter(confirmed > 100 & country != "Diamond Princess" & country != 'China') %>%
  group_by(country) %>%
  filter(n() > 11) %>%
  arrange(country, date) %>%
  # fix values for March 12 by averaging the days before and after
  mutate(confirmed = ifelse(date == "2020-03-12", 
                            lag(confirmed)+(lead(confirmed)-lag(confirmed))/2,
                            confirmed)) %>%
  arrange(country, date) %>%
  # compute growth rate for past 5 days
  mutate(daysago = lag(confirmed, n=5)) %>% 
  mutate(growth = 100* (( confirmed / daysago )^(1/5) - 1)) %>%
  # and take a 3-day rolling average from it
  mutate(growth = (rollapplyr(growth, 3, mean, fill=NA))) %>% 
  select(country, date, growth) %>% 
  pivot_wider(names_from = "country", values_from = "growth") %>%
  filter(date > '2020-02-29') %>%
  arrange(desc(date)) %>%
  write_csv('growth-rates.csv')