felixgolcher/corona_wp.R

## corona_wp.csv

          
            date
            suspected
            confirmed
            deaths

            
              2020-01-10
              NA
              41
              1

            
              2020-01-11
              NA
              41
              1

            
              2020-01-12
              NA
              41
              1

            
              2020-01-13
              NA
              41
              1

            
              2020-01-15
              NA
              NA
              2

            
              2020-01-16
              NA
              45
              NA

            
              2020-01-17
              NA
              62
              2

            
              2020-01-18
              NA
              121
              NA

            
              2020-01-19
              NA
              198
              3

            
              2020-01-20
              54
              291
              6

            
              2020-01-21
              37
              440
              9

            
              2020-01-22
              257
              571
              17

            
              2020-01-23
              1072
              830
              25

            
              2020-01-24
              1965
              1287
              41

            
              2020-01-25
              2684
              1975
              56

            
              2020-01-26
              5794
              2744
              80

            
              2020-01-27
              6973
              4515
              106

            
              2020-01-28
              9239
              5974
              132

            
              2020-01-29
              12167
              7711
              170

            
              2020-01-30
              15238
              9692
              213

            
              2020-01-31
              17988
              11791
              259

            
              2020-02-01
              19544
              14380
              304

            
              2020-02-02
              21558
              17205
              361

            
              2020-02-03
              23214
              20440
              425

            
              2020-02-04
              23260
              24324
              490

            
              2020-02-05
              24702
              28018
              563

            
              2020-02-06
              26359
              31161
              637

            
              2020-02-07
              27657
              34568
              722

            
              2020-02-08
              28942
              37198
              811

## corona_wp.R
library(tidyverse)
library(ggplot2)
# using csv now
# library(readODS)

# height of figs
mh <- 3
# width
mw <- mh*1.5

# data (manually) gathered from
# https://en.wikipedia.org/wiki/Timeline_of_the_2019%E2%80%9320_Wuhan_coronavirus_outbreak

### converting the orig ods file.
# cwp <- read_ods("corona_wp.ods")
# write_csv(cwp, "corona_wp.csv")
cwp <- read_csv("corona_wp.csv")

# even linear display of confirmed shows a downard curvature.
(ggplot(cwp, aes(date, confirmed))+
  geom_point()+
    ggtitle("Confirmed cases in linear scale")+
  theme_minimal()->gg1)
ggsave("confirmed_lin.png", width=mw, height = mh)

# not an exponential development for quite a while:
(gg1 + scale_y_log10("confirmed (log scale)")+
    ggtitle("Confirmed cases in log scale") -> gg1a)
ggsave("confirmed_log.png", width=mw, height = mh)

# deaths more regular, more exp.:
(ggplot(cwp, aes(date, deaths))+
  geom_point()+
    ggtitle("Deaths in linear scale")+
  theme_minimal()-> gg2)
ggsave("deaths_lin.png", width=mw, height = mh)

# but not really exponential either, even if you can decide to call stretches of
# it exponential.
gg2 + scale_y_log10("deaths (log scale)")+
  ggtitle("Deaths in log scale")
ggsave("deaths_log.png", width=mw, height = mh)

(cwp %>%
  pivot_longer(-date, names_to = "description", values_to = "count")->cwp.long) %>%
  filter(description != "suspected") %>%
  ## suspected cases are current, not cumulative.
  ggplot(aes(date, count, col=description))+
  geom_point()+
  scale_y_log10()+
  theme_minimal()
ggsave("both_log.png", width=mw, height = mh)

## there was a strange blip in suspected cases on feb 03/04.
ggplot(cwp, aes(date, suspected))+
  geom_point()

## deaths by confirmed went down for a long while, and then up again, but that
## is to be expected as the number of confirmed cases grows slower.
ggplot(cwp, aes(date, deaths/confirmed, size=confirmed))+
  geom_point(alpha=.7)
ggsave("frac.png", width=mw, height = mh)

## the relative daily increase goes consistently down.
cwp %>% mutate(inc.perc = confirmed/lag(confirmed)-1) %>%
  ggplot(aes(date, inc.perc, size=confirmed))+
  geom_point(alpha=.7)+
  ggtitle("daily relative increase in confirmed cases")+
  scale_y_continuous("increase in percent")
ggsave("inc.png", width=mw, height = mh)

# the _absolute_ daily increase shows a somewhat strange behaviour. It could be
# explained by a genuine success of the chinese efforts to combat the virus or
# on limited testing capacities. The latter explanation cannot easily cope with
# the general downward trend over the last few days.
cwp %>% mutate(inc.abs = confirmed-lag(confirmed)) %>%
  ggplot(aes(date, inc.abs, size=confirmed))+
  geom_point(alpha=.7)+
  ggtitle("daily absolute increase in confirmed cases")+
  scale_y_continuous("absolute increase")
ggsave("incabs.png", width=mw, height = mh)

## the daily increase in deaths does not follow the pattern of confirmed cases
## (yet?), but that is not to be expected either.
cwp %>% mutate(inc.abs = deaths-lag(deaths)) %>%
  ggplot(aes(date, inc.abs, size=confirmed))+
  ggtitle("daily absolute increase in deaths")+
  geom_point(alpha=.7)+
  scale_y_continuous("absolute increase (deaths)")
ggsave("incabs-death.png", width=mw, height = mh)
date	suspected	confirmed	deaths
2020-01-10	NA	41	1
2020-01-11	NA	41	1
2020-01-12	NA	41	1
2020-01-13	NA	41	1
2020-01-15	NA	NA	2
2020-01-16	NA	45	NA
2020-01-17	NA	62	2
2020-01-18	NA	121	NA
2020-01-19	NA	198	3
2020-01-20	54	291	6
2020-01-21	37	440	9
2020-01-22	257	571	17
2020-01-23	1072	830	25
2020-01-24	1965	1287	41
2020-01-25	2684	1975	56
2020-01-26	5794	2744	80
2020-01-27	6973	4515	106
2020-01-28	9239	5974	132
2020-01-29	12167	7711	170
2020-01-30	15238	9692	213
2020-01-31	17988	11791	259
2020-02-01	19544	14380	304
2020-02-02	21558	17205	361
2020-02-03	23214	20440	425
2020-02-04	23260	24324	490
2020-02-05	24702	28018	563
2020-02-06	26359	31161	637
2020-02-07	27657	34568	722
2020-02-08	28942	37198	811
	library(tidyverse)
	library(ggplot2)
	# using csv now
	# library(readODS)

	# height of figs
	mh <- 3
	# width
	mw <- mh*1.5

	# data (manually) gathered from
	# https://en.wikipedia.org/wiki/Timeline_of_the_2019%E2%80%9320_Wuhan_coronavirus_outbreak

	### converting the orig ods file.
	# cwp <- read_ods("corona_wp.ods")
	# write_csv(cwp, "corona_wp.csv")
	cwp <- read_csv("corona_wp.csv")

	# even linear display of confirmed shows a downard curvature.
	(ggplot(cwp, aes(date, confirmed))+
	geom_point()+
	ggtitle("Confirmed cases in linear scale")+
	theme_minimal()->gg1)
	ggsave("confirmed_lin.png", width=mw, height = mh)

	# not an exponential development for quite a while:
	(gg1 + scale_y_log10("confirmed (log scale)")+
	ggtitle("Confirmed cases in log scale") -> gg1a)
	ggsave("confirmed_log.png", width=mw, height = mh)

	# deaths more regular, more exp.:
	(ggplot(cwp, aes(date, deaths))+
	geom_point()+
	ggtitle("Deaths in linear scale")+
	theme_minimal()-> gg2)
	ggsave("deaths_lin.png", width=mw, height = mh)

	# but not really exponential either, even if you can decide to call stretches of
	# it exponential.
	gg2 + scale_y_log10("deaths (log scale)")+
	ggtitle("Deaths in log scale")
	ggsave("deaths_log.png", width=mw, height = mh)

	(cwp %>%
	pivot_longer(-date, names_to = "description", values_to = "count")->cwp.long) %>%
	filter(description != "suspected") %>%
	## suspected cases are current, not cumulative.
	ggplot(aes(date, count, col=description))+
	geom_point()+
	scale_y_log10()+
	theme_minimal()
	ggsave("both_log.png", width=mw, height = mh)

	## there was a strange blip in suspected cases on feb 03/04.
	ggplot(cwp, aes(date, suspected))+
	geom_point()

	## deaths by confirmed went down for a long while, and then up again, but that
	## is to be expected as the number of confirmed cases grows slower.
	ggplot(cwp, aes(date, deaths/confirmed, size=confirmed))+
	geom_point(alpha=.7)
	ggsave("frac.png", width=mw, height = mh)

	## the relative daily increase goes consistently down.
	cwp %>% mutate(inc.perc = confirmed/lag(confirmed)-1) %>%
	ggplot(aes(date, inc.perc, size=confirmed))+
	geom_point(alpha=.7)+
	ggtitle("daily relative increase in confirmed cases")+
	scale_y_continuous("increase in percent")
	ggsave("inc.png", width=mw, height = mh)

	# the _absolute_ daily increase shows a somewhat strange behaviour. It could be
	# explained by a genuine success of the chinese efforts to combat the virus or
	# on limited testing capacities. The latter explanation cannot easily cope with
	# the general downward trend over the last few days.
	cwp %>% mutate(inc.abs = confirmed-lag(confirmed)) %>%
	ggplot(aes(date, inc.abs, size=confirmed))+
	geom_point(alpha=.7)+
	ggtitle("daily absolute increase in confirmed cases")+
	scale_y_continuous("absolute increase")
	ggsave("incabs.png", width=mw, height = mh)

	## the daily increase in deaths does not follow the pattern of confirmed cases
	## (yet?), but that is not to be expected either.
	cwp %>% mutate(inc.abs = deaths-lag(deaths)) %>%
	ggplot(aes(date, inc.abs, size=confirmed))+
	ggtitle("daily absolute increase in deaths")+
	geom_point(alpha=.7)+
	scale_y_continuous("absolute increase (deaths)")
	ggsave("incabs-death.png", width=mw, height = mh)