-
Notifications
You must be signed in to change notification settings - Fork 2
/
0195-covid-symptoms-blogpost.R
174 lines (154 loc) · 8.17 KB
/
0195-covid-symptoms-blogpost.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
# This script is for producing static images for a blog post on Free Range Statistics
#
# This script isn't self sufficient, you need to have loaded packages nad hte data first.
# Run this script from build.R
#
# Peter Ellis 4 October 2020
the_caption <- "University of Maryland Global Symptom Survey of Facebook users, analysed at http://freerangestats.info"
the_countries <- c("Australia", "Brazil", "Singapore", "Italy", "New Zealand", "United Kingdom")
# colours for male and female used by Washington Post 2017; see https://blog.datawrapper.de/gendercolor/
sex_cols <- c(Male = "#F4BA3B", Female = "#730B6D")
#==============Australia==========
oz_regions_5 <- raw_data_regions %>%
filter(country_agg == "Australia" & gender == "overall" & age_bucket == "overall") %>%
mutate(state = fct_lump(region_agg, 5,w = total_responses)) %>%
group_by(state, date) %>%
# aggregate, weighted by total responses, to combine Tasmania and ACT:
summarise(pct_cli_weighted = weighted.mean(pct_cli_weighted, w = total_responses) / 100,
pct_feel_depressed_none_time_weighted = weighted.mean(
pct_feel_depressed_none_time_weighted, w = total_responses) / 100,
pct_wear_mask_none_time_weighted = weighted.mean(
pct_wear_mask_none_time_weighted, w = total_responses, na.rm = TRUE) / 100,
total_responses = sum(total_responses)
) %>%
ungroup() %>%
mutate(state = fct_reorder(state, -total_responses, .fun = sum))
#-----------------Symptoms by state and time--------
p1 <- oz_regions_5 %>%
ggplot(aes(x = date, y = pct_cli_weighted)) +
geom_point(aes(size = total_responses)) +
geom_smooth(se = FALSE, span = 1/4, method = "loess", formula = "y ~ x") +
labs(subtitle = str_wrap("Percentage of population showing Covid-like illness symptoms. Victoria
is not higher than other states; nor is there a spike in Victoria in July or August,
even though in both cases there should be two orders of magnitude of variation.", 120),
title = "Survey-reported Covid-like symptoms in Australia **do not** follow known spatial or temporal trends",
y = "'Have symptoms of Covid-like illness'",
x = "Date in 2020",
size = "Sample size:",
caption = the_caption) +
coord_cartesian(ylim = c(0, 0.08)) +
scale_size_area(max_size = 5, label = comma) +
scale_y_continuous(label = percent_format(accuracy = 1)) +
facet_wrap(~state, scales = "fixed")
#---------------------Mask usage------------------
p2 <- oz_regions_5 %>%
ggplot(aes(x = date, y = pct_wear_mask_none_time_weighted)) +
geom_point(aes(size = total_responses)) +
geom_smooth(se = FALSE, span = 1/4, method = "loess", formula = "y ~ x") +
labs(subtitle = str_wrap("Percentage of population reporting they never wear a mask in Victoria
declined rapidly to zero when mask-wearing became mandatory. Other states with lesser
or no outbreak report much less mask wearing, reflecting what we know to be reality.", 120),
title = "Survey-reported wearing of masks **does** follow known spatial and temporal trends",
y = "'Never wear a mask'",
x = "Date in 2020",
size = "Sample size:",
caption = the_caption) +
coord_cartesian(ylim = c(0, 1)) +
scale_size_area(max_size = 5, label = comma) +
scale_y_continuous(label = percent_format(accuracy = 1)) +
facet_wrap(~state, scales = "fixed")
#---------------------Depression------------------
p3 <- oz_regions_5 %>%
ggplot(aes(x = date, y = pct_feel_depressed_none_time_weighted)) +
geom_point(aes(size = total_responses)) +
geom_smooth(se = FALSE, span = 1/4, method = "loess", formula = "y ~ x") +
labs(subtitle = str_wrap("Percentage of population reporting they feel depressed 'none of the time'
markedly drops in the period of the serious Covid-19 outbreak and response from June
to September.", 120),
title = "Survey-reported levels of depression **are** related to Covid-19 incidence and response",
y = "'Feel depressed none of the time'",
x = "Date in 2020",
size = "Sample size:",
caption = the_caption) +
scale_size_area(max_size = 5, label = comma) +
scale_y_continuous(label = percent_format(accuracy = 1)) +
facet_wrap(~state, scales = "fixed")
#==============six countries==========
countries <- raw_data_countries %>%
filter(country_agg %in% the_countries) %>%
filter(age_bucket == "overall") %>%
filter(gender %in% c("female", "male")) %>%
mutate(gender = str_to_title(gender))
#-----------------Symptoms by state and time--------
p4 <- countries %>%
ggplot(aes(x = date, y = pct_cli_weighted / 100, colour = gender)) +
geom_point(aes(size = total_responses), alpha = 0.5) +
geom_smooth(se = FALSE, span = 1/4, method = "loess", formula = "y ~ x") +
labs(title = "Survey-reported Covid-like symptoms in selected countries",
subtitle = str_wrap("Brazil has relatively high level of reported symptoms as might
be expected giving Covid prevalence, but so does New Zealand. UK has seen a recent
increase, which
is as expected (perhaps). Australia's trend in reported symptoms does not follow
trend in incidence.", 120),
y = "'Have symptoms of Covid-like illness'",
x = "Date in 2020",
size = "Sample size:",
caption = the_caption,
colour = "") +
scale_size_area(max_size = 5, label = comma) +
scale_y_continuous(label = percent_format(accuracy = 1)) +
scale_colour_manual(values = sex_cols) +
facet_wrap(~country_agg, scales = "fixed")
#---------------------Mask usage------------------
p5 <- countries %>%
ggplot(aes(x = date, y = pct_wear_mask_none_time_weighted / 100, colour = gender)) +
geom_point(aes(size = total_responses), alpha = 0.5) +
geom_smooth(se = FALSE, span = 1/4, method = "loess", formula = "y ~ x") +
labs(title = "Survey-reported wearing of masks in selected countries",
subtitle = str_wrap("Reported mask wearing follows clear, well-defined patterns
that fit with expectations of reality. The response to outbreaks
in Australia and New Zealand are clear, as is the gradual
adoption over time in the United Kingdom.", 120),
y = "'Never wear a mask'",
x = "Date in 2020",
size = "Sample size:",
caption = the_caption,
colour = "") +
coord_cartesian(ylim = c(0, 1)) +
scale_size_area(max_size = 5, label = comma) +
scale_y_continuous(label = percent_format(accuracy = 1)) +
scale_colour_manual(values = sex_cols) +
facet_wrap(~country_agg, scales = "fixed")
#---------------------Depression------------------
p6 <- countries %>%
ggplot(aes(x = date, y = pct_feel_depressed_none_time_weighted / 100, colour = gender)) +
geom_point(aes(size = total_responses), alpha = 0.5) +
geom_smooth(se = FALSE, span = 1/4, method = "loess", formula = "y ~ x") +
labs(subtitle = str_wrap("Italians' self-reported depression is improving over time,
while in Australia it took a hit around July. In New Zealand,
women rather than men reported more depression at the time
of the outbreak around August.", 120),
title = "Survey-reported levels of depression in selected countries",
y = "'Feel depressed none of the time'",
x = "Date in 2020",
size = "Sample size:",
caption = the_caption,
colour = "") +
scale_size_area(max_size = 5, label = comma) +
scale_colour_manual(values = sex_cols) +
scale_y_continuous(label = percent_format(accuracy = 1)) +
facet_wrap(~country_agg, scales = "fixed")
#===============saving===========
w <- 10
h <- 6
svg_png(p1, "output/0195-states-cli", w = w, h = h)
svg_png(p2, "output/0195-states-masks", w = w, h = h)
svg_png(p3, "output/0195-states-dep", w = w, h = h)
svg_png(p4, "output/0195-countries-cli", w = w, h = h)
svg_png(p5, "output/0195-countries-masks", w = w, h = h)
svg_png(p6, "output/0195-countries-dep", w = w, h = h)
#---------Copy files over-----------
files <- list.files("output", pattern = "0195")
file.copy(glue("output/{files}"),
glue("~/blog/ellisp.github.io/img/{files}"),
overwrite = TRUE)