kyle.belanger/DHSC-Capstone
1
0
Fork 0
Code Issues 1 Pull requests Projects Releases Packages Wiki Activity Actions

Update 1-data-exploration.R

Browse source
This commit is contained in:
Kyle Belanger 2023-01-24 13:58:35 -05:00
parent 538f9f76e5
commit bbe5cd1ae8
1 changed files with 4 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
ML/1-data-exploration.R
View file

@ -30,7 +30,6 @@ ds0 <- readr$read_rds(here("ML","data-unshared","ds_final.RDS"))
ds1 <- ds0 %>% ds1 <- ds0 %>%
dplyr$select(-subject_id, -charttime) %>%
dplyr$mutate(dplyr$across( dplyr$mutate(dplyr$across(
ft4_dia ft4_dia
, ~factor(., levels = c("Hypo", "Non-Hypo","Hyper", "Non-Hyper") , ~factor(., levels = c("Hypo", "Non-Hypo","Hyper", "Non-Hyper")
@ -61,6 +60,7 @@ ds_recode <- ds1 %>%
#summary Table #summary Table
summary_tbl <- ds1 %>% summary_tbl <- ds1 %>%
dplyr$select(-subject_id, -charttime) %>%
gtsummary$tbl_summary( gtsummary$tbl_summary(
by = ft4_dia by = ft4_dia
,missing = "no" ,missing = "no"
@ -88,7 +88,7 @@ summary_tbl <- ds1 %>%
# correlation plot # correlation plot
corr_plot <- cor( corr_plot <- cor(
ds1 %>% dplyr$select(-gender,-ft4_dia) ds1 %>% dplyr$select(-gender,-ft4_dia, -subject_id, -charttime)
,use = "complete.obs" ,use = "complete.obs"
) %>% ) %>%
corrplot::corrplot(method = "number", type = "lower", tl.col = "black", tl.srt = 45 corrplot::corrplot(method = "number", type = "lower", tl.col = "black", tl.srt = 45
@ -105,7 +105,7 @@ dev.off()
#quick recode of gender, will still do recoding during feature engineering #quick recode of gender, will still do recoding during feature engineering
g1 <- ds1 %>% g1 <- ds1 %>%
dplyr$select(-gender, -ft4_dia) %>% dplyr$select(-gender,-ft4_dia, -subject_id, -charttime) %>%
tidyr$pivot_longer(cols = dplyr$everything()) %>% tidyr$pivot_longer(cols = dplyr$everything()) %>%
ggplot(aes(x = value)) + ggplot(aes(x = value)) +
gp2$geom_histogram(na.rm = TRUE) + gp2$geom_histogram(na.rm = TRUE) +
@ -128,7 +128,7 @@ gp2$ggsave(
# this takes a bit to load. No discernible patterns in the data # this takes a bit to load. No discernible patterns in the data
g2 <- ds_recode %>% g2 <- ds_recode %>%
dplyr$select(-gender) %>% dplyr$select(-gender, -subject_id, -charttime) %>%
dplyr$mutate(dplyr$across(-ft4_dia, log)) %>% dplyr$mutate(dplyr$across(-ft4_dia, log)) %>%
tidyr$pivot_longer(cols = !ft4_dia) %>% tidyr$pivot_longer(cols = !ft4_dia) %>%
ggplot(aes(x = factor(ft4_dia), y = value, fill = factor(ft4_dia))) + ggplot(aes(x = factor(ft4_dia), y = value, fill = factor(ft4_dia))) +