-
Notifications
You must be signed in to change notification settings - Fork 1
/
ohidata.Rmd
310 lines (275 loc) · 9.34 KB
/
ohidata.Rmd
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
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
---
title: "OHI Metrics"
output: html_document
---
```{r include=FALSE}
knitr::opts_chunk$set(echo = TRUE, message = FALSE)
```
```{r}
library(tidyverse)
library(ohiCovidMetrics)
library(readxl)
source("R/shape_data.R")
source("R/process_metrics.R")
```
Carl, as you begin to use more sensitive data, it would be useful to have APIs that allow user to specify the data origin and possibly the extraction tool. For instance, you are starting to use ODBC (Oracle), but I think we get copies of data from DHS into our SILO protected area as flat files. This is down the way (weeks), and we can work together on how to do this.
This repository <https://github.com/byandell/gating>, which uses the [ohiCovidMetrics](https://github.com/carlbfrederick/ohiCovidMetrics) package, is where Yandell is exploring ideas.
Please note the last image, which highlights counties that have had >5% positive tests during one of the past 4 weeks.
Notice that I have modified versions of a couple of your routines. Key differences:
- removed ".data$" for better readability
- modified shape_data as generalization of shape_case_data:
+ Took out pivot_wider() to keep longer format (go wide for plotting only)
+ Added "weeks" option to look at more than 2 weeks
+ Added test and death data
- modified process_metrics as generalization of process_confirmed_cases
+ Handle long format
+ Handle all count types at once
Part of my philosophy is to have more general routines when it makes sense to process similar data, here case, death, test. Note that I ran data through your metrics without adjustment, but down the road the guidelines for deaths or %pos could be added. Just an idea.
Just a heads up that [ohiCovidMetrics](https://github.com/carlbfrederick/ohiCovidMetrics) requires R 4.0 and the `sf` and `ROBDC` packages. Upgrade to a major R version, requires reinstalling all your packages, such as `tidyverse`.
## Data setup
```{r pull-hist-table}
hdt_raw <- pull_histTable(end_date = "2020-06-17")
```
```{r shape_data}
weeks <- 10
hdt_clean <- shape_data(hdt_raw, weeks = weeks)
```
```{r eval=FALSE}
hdt_out <- process_metrics(
hdt_clean %>%
filter(weeknum <= 2) %>%
pivot_wider(id_cols = c("fips", "geo_type", "geo_name", "pop_2018", "count_type"),
values_from = c("weekly", "week_end", "pct_pos"),
names_from = "weeknum"))
```
```{r}
hdt_clean_pct <- hdt_clean %>%
filter(weeknum <= 4) %>%
group_by(fips, geo_type, geo_name, pop_2018, count_type) %>%
summarize(maxpospct = max(pct_pos),
minpospct = min(pct_pos),
curpospct = pct_pos[1],
maxcount = max(weekly),
mincount = min(weekly),
curcount = weekly[1]) %>%
ungroup()
```
## Pull out counties only
```{r}
hdt_co <-
left_join(
hdt_raw %>%
filter(geo_type == "County") %>%
rename(county = "geo_name"),
county_data,
by = c("fips", "county", "pop_2018"))
```
```{r eval=FALSE}
write.csv(distinct(hdt_co, county, dph_region, herc_region), file="herc.csv", row.names=FALSE)
```
```{r}
hdt_co_clean <-
left_join(
hdt_clean %>%
filter(geo_type == "County") %>%
rename(county = "geo_name"),
county_data,
by = c("fips", "county", "pop_2018"))
```
## Plots of Raw data
```{r}
ggplot(hdt_raw %>%
filter(geo_type == "HERC Region")) +
aes(post_date, case_cum, col = geo_name) +
geom_line() +
xlab("Date") +
ylab("Cumulative Cases") +
scale_y_log10() +
ggtitle("Cumulative Counts over Time")
```
```{r}
ggplot(hdt_raw %>%
filter(geo_type == "HERC Region")) +
aes(test_cum, case_cum, col = geo_name) +
geom_abline(slope = 1, intercept = 0, col = "grey") +
geom_line() +
xlab("Cumulative Tests") +
ylab("Cumulative Cases") +
scale_x_log10() +
scale_y_log10() +
ggtitle("Cumulative Testing vs Cases")
```
```{r}
ggplot(hdt_raw %>%
filter(geo_type == "HERC Region",
max(post_date) - post_date <= 14)) +
aes(post_date, case_daily, col = geo_name) +
geom_line() +
xlab("Date") +
ylab("New Cases") +
scale_y_log10() +
ggtitle("New Cases by HERC Regions in Past 2 Weeks")
```
```{r}
ggplot(hdt_co %>%
filter(max(post_date) - post_date <= 14)) +
aes(post_date, case_daily, col = county) +
geom_line() +
facet_wrap(~ herc_region, scales = "free_y") +
theme(legend.position = "none") +
xlab("Date") +
ylab("New Cases") +
scale_y_log10() +
ggtitle("New Cases by County in Past 2 Weeks")
```
```{r}
ggplot(hdt_co %>%
filter(max(post_date) - post_date <= 14)) +
aes(post_date, test_daily, col = county) +
geom_line() +
facet_wrap(~ herc_region, scales = "free_y") +
xlab("Date") +
ylab("New Cases") +
theme(legend.position = "none") +
scale_y_log10() +
ggtitle("New Testing by County in Past 2 Weeks")
```
## Plots of Clean Data
```{r}
ggplot(hdt_co_clean %>%
filter(count_type == "case",
weeknum < 3) %>%
pivot_wider(id_cols = c("county", "pop_2018", "herc_region"),
values_from = c("weekly", "week_end", "pct_pos"),
names_from = "weeknum") %>%
filter(weekly_2 != weekly_1)) +
aes(weekly_1, weekly_2 - weekly_1, col = county) +
geom_hline(yintercept = 0, col = "grey") +
geom_point() +
facet_wrap(~herc_region, scales = "free") +
theme(legend.position = "none") +
xlab("Current Week Count") +
ylab("Change from Last Week") +
ggtitle("Counties with Changes") +
scale_x_log10()
```
```{r warning=FALSE}
ggplot(hdt_co_clean %>%
filter(count_type == "case")) +
aes(weekly, pct_pos, col = county, label = weeknum) +
geom_hline(yintercept = 0, col = "grey") +
geom_text() +
geom_path() +
facet_wrap(~ herc_region, scales = "free") +
theme(legend.position = "none") +
xlab("Weekly Confirmed Cases") +
ylab("Percent Positive") +
scale_x_log10() +
ggtitle(paste("Cases by % Positive over Past", weeks, "Weeks"))
```
```{r warning=FALSE}
ggplot(hdt_co_clean %>%
filter(count_type != "death")) +
aes(weekly, pct_pos, col = county, label = weeknum) +
geom_hline(yintercept = 0, col = "grey") +
geom_path() +
geom_text() +
facet_grid(herc_region ~ count_type, scales = "free") +
theme(legend.position = "none") +
xlab("Weekly Count") +
ylab("Percent Positive") +
scale_x_log10() +
ggtitle(paste("Cases and Tests by % Positive over Past", weeks, "Weeks"))
```
```{r}
(co_week4_pct5 <- left_join(
hdt_co_clean,
hdt_clean_pct %>%
rename(county = "geo_name"),
by = c("fips","geo_type","county","pop_2018","count_type")) %>%
filter(count_type == "case",
weeknum <= 4,
maxpospct >= 5)) %>%
filter(count_type == "case") %>%
select(county, weeknum, weekly, pct_pos, herc_region) %>%
rename(case = "weekly", pct = "pct_pos") %>%
pivot_wider(id_cols = c("county","herc_region"),
values_from = c("case", "pct"),
names_from = "weeknum") %>%
arrange(herc_region, county)
```
```{r warning=FALSE}
ggplot(co_week4_pct5 %>%
mutate(HighPos = (curpospct < 5))) +
aes(weekly, pct_pos, col = HighPos, group = county, label = weeknum) +
geom_hline(yintercept = 0, col = "grey") +
geom_path() +
geom_text() +
facet_wrap(~ herc_region, scales = "free") +
theme(legend.position = "none") +
xlab("Weekly Count") +
ylab("Percent Positive") +
ggtitle("Last 4 Weeks for Counties with >5% Positive Tests") +
scale_x_log10()
```
```{r}
co_week4_pct5l <- left_join(
hdt_co_clean,
hdt_clean_pct %>%
rename(county = "geo_name"),
by = c("fips","geo_type","county","pop_2018","count_type")) %>%
filter(count_type == "case",
weeknum <= 4,
maxpospct < 5,
maxcount > 10)
```
```{r warning=FALSE}
ggplot(co_week4_pct5l %>%
mutate(HighCase = (curcount < 10))) +
aes(weekly, pct_pos, col = HighCase, group = county, label = weeknum) +
geom_hline(yintercept = 0, col = "grey") +
geom_path() +
geom_text() +
facet_wrap(~ herc_region, scales = "free") +
theme(legend.position = "none") +
xlab("Weekly Count") +
ylab("Percent Positive") +
ggtitle("Last 4 Weeks for Counties with >10 Confirmed Cases") +
scale_x_log10()
```
```{r}
co_week4_pct5b <- left_join(
hdt_co_clean,
hdt_clean_pct %>%
rename(county = "geo_name"),
by = c("fips","geo_type","county","pop_2018","count_type")) %>%
filter(count_type == "case",
weeknum <= 4,
maxpospct >= 5 | (maxpospct < 5 & maxcount > 10))
```
```{r warning=FALSE}
ggplot(co_week4_pct5b %>%
mutate(HighPct = factor((curpospct >= 5) + 2 * (curpospct < 5 & curcount >= 10)))) +
aes(weekly, pct_pos, col = HighPct, group = county, label = weeknum) +
geom_hline(yintercept = 0, col = "grey") +
geom_path() +
geom_text() +
facet_wrap(~ herc_region, scales = "free") +
theme(legend.position = "none") +
xlab("Weekly Count") +
ylab("Percent Positive") +
ggtitle("Last 4 Weeks for Counties with >5% Positive or >10 Confirmed Cases") +
scale_x_log10()
```
```{r}
co_week4_pct5b %>%
filter(count_type == "case") %>%
select(county, weeknum, weekly, pct_pos, herc_region) %>%
rename(case = "weekly", pct = "pct_pos") %>%
pivot_wider(id_cols = c("county","herc_region"),
values_from = c("case", "pct"),
names_from = "weeknum") %>%
arrange(herc_region, desc(pct_1)) %>%
filter(pct_1 >= 5 | (pct_1 < 5 & case_1 >= 10)) %>%
knitr::kable()
```