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 2023-01-22 08:11:50 -05:00
parent 66f213bc5a
commit 4486218a6e
1 changed files with 33 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

43
ML/1-data-exploration.R
View file

@ -38,6 +38,24 @@ ds1 <- ds0 %>%
) )
) )
ds_recode <- ds1 %>%
dplyr$mutate(
dplyr$across(
gender
,~dplyr$recode(.,"M" = 1, "F" = 2)
)
,dplyr$across(
ft4_dia
,~dplyr$recode(.
,"Hypo" = 1
,"Non-Hypo" = 2
,"Normal TSH" = 3
,"Hyper" = 4
,"Non-Hyper" = 5
)
)
)
# basic visualization ----------------------------------------------------- # basic visualization -----------------------------------------------------
@ -70,39 +88,44 @@ summary_tbl <- ds1 %>%
# correlation plot # correlation plot
ds_corr <- cor(ds_high_tsh %>% ds_corr <- cor(ds_recode,use = "complete.obs")
dplyr$mutate(dplyr$across(gender, ~dplyr$recode(.,M = 1, F = 2)))
,use = "complete.obs")
#code for saving corr plot #code for saving corr plot
png(here("figures","corrplot_high.png"), type = 'cairo') png(here("figures","corrplot.png"), type = 'cairo')
corrplot::corrplot(ds_corr, method = "number") corrplot::corrplot(ds_corr, method = "number")
dev.off() 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 <- ds_high_tsh %>% g1 <- ds1 %>%
dplyr$mutate(dplyr$across(gender, ~dplyr$recode(.,M = 1, F = 2))) %>% dplyr$select(-gender, -ft4_dia) %>%
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) +
gp2$facet_wrap(~name, scales = "free") gp2$facet_wrap(~name, scales = "free") +
gp2$theme_bw() +
gp2$labs(
x = NULL
,y = NULL
)
g1 g1
# this takes a bit to load. No discernable paterns in the data # this takes a bit to load. No discernable paterns in the data
g2 <- ds_high_tsh %>% g2 <- ds_recode %>%
dplyr$select(-gender) %>% dplyr$select(-gender) %>%
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))) +
gp2$geom_boxplot(outlier.shape = NA, na.rm = TRUE) + gp2$geom_boxplot(outlier.shape = NA, na.rm = TRUE) +
gp2$geom_jitter(size=.7, width=.1, alpha=.5, na.rm = TRUE) + gp2$geom_jitter(size=.7, width=.1, alpha=.5, na.rm = TRUE) +
gp2$facet_wrap(~name, scales = "free") gp2$facet_wrap(~name, scales = "free") +
gp2$theme_bw() +
gp2$scale_fill_brewer(palette = "Greys")
g2 g2