-
Notifications
You must be signed in to change notification settings - Fork 0
/
eda.Rmd
302 lines (251 loc) · 9.22 KB
/
eda.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
---
title: "Exploratory Data Analysis of ICU Patient Dataset"
output: html_document
date: "2024-07-31"
---
```{r setup, include=FALSE}
# Load necessary libraries
library(ggplot2)
library(patchwork)
library(dplyr)
library(janitor)
```
```{r}
# Load the dataset
merged_df <- read.csv("merged_df_NEW.csv")
```
# Data Cleaning
```{r}
# Clean the column names and remove empty columns.
merged_df_clean <- merged_df %>%
clean_names() %>%
remove_empty("cols")
merged_df_clean = subset(merged_df_clean, select = -c(deathtime, survival_time, admittime, group, subject_id))
```
# Missing Data Analysis
```{r, fig.width 10}
# Filter the dataframe to include only columns with missing values.
merged_df_filtered <- merged_df_clean %>%
select_if(~ any(is.na(.)))
# Summarize the missing data for each column.
missing_data <- merged_df_filtered %>%
summarise_all(~sum(is.na(.))) %>%
gather(key = "variable", value = "missing_count") %>%
mutate(
total_count = nrow(merged_df_filtered),
missing_percent = (missing_count / total_count) * 100
)
# Create a vector with more readable variable names.
new_names <- c(
pco2 = "PCO2",
ph = "pH",
basophils = "Basophils",
lactic_acid = "Lactic Acid",
bmi = "BMI",
creatine_kinase = "Creatine Kinase",
los = "LOS",
lymphocyte = "Lymphocyte",
neutrophils = "Neutrophils",
language = "Language",
urine_output = "Urine Output",
marital_status = "Marital Status",
pt = "PT",
inr = "INR",
temperature = "Temperature",
glucose = "Glucose",
systolic_blood_pressure = "Systolic Blood Pressure",
diastolic_blood_pressure = "Diastolic Blood Pressure",
sp_o2 = "SpO2",
respiratory_rate = "Respiratory Rate",
heart_rate = "Heart Rate",
blood_calcium = "Blood Calcium"
)
# Plot the missing data overview with bar plot.
ggplot(missing_data, aes(x = reorder(variable, -missing_count), y = missing_count)) +
geom_bar(stat = "identity", fill = "steelblue", width = 0.8) +
geom_text(aes(label = paste0(round(missing_percent, 1), "%")),
vjust = -0.5,
color = "black",
size = 3, # Adjust size to minimize text
nudge_y = 5) +
labs(
title = "Missing Data Overview",
x = "Variables",
y = "Number of Missing Values"
) +
scale_x_discrete(labels = new_names) + # Apply the new names
theme(axis.text.x = element_text(angle = 45, hjust = 1))
```
# Data Visualization
```{r}
# Define a function to create histograms for specified variables and their units.
create_histogram <- function(merged_df, variable, units) {
mean_alive <- mean(merged_df[merged_df$outcome == "0", variable], na.rm = TRUE)
mean_dead <- mean(merged_df[merged_df$outcome == "1", variable], na.rm = TRUE)
p <- ggplot(merged_df, aes_string(x = variable, fill = "factor(outcome)")) +
geom_histogram(aes(y = ..density..), position = "identity", alpha = 0.5, bins = 20) +
geom_vline(xintercept = mean_alive, col = "blue", lwd = 1, linetype = "dotted") +
geom_vline(xintercept = mean_dead, col = "green", lwd = 1, linetype = "dotted") +
scale_fill_manual(values = c("blue", "green"), labels = c("Alive", "Dead")) +
labs(title = paste("", gsub("\\.", " ", variable), "by Outcome"),
x = paste(gsub("\\.", " ", variable), ifelse(variable == "PH", "", paste("(", units, ")"))),
y = "Density",
fill = "Outcome") +
theme_minimal() +
theme(
plot.title = element_text(size = 12, face = "bold"),
axis.title = element_text(size = 10),
axis.text = element_text(size = 8)
)
return(p)
}
# Create a list of variables and their units for the histogram creation.
variables <- list(
"Blood.calcium" = "mEq/L",
"Bicarbonate" = "mg/dL",
"PH" = "hydrogen ion concentration",
"Urea.nitrogen" = "mg/dL"
)
# Apply the create_histogram function to each variable and combine the plots.
plots <- lapply(names(variables), function(var) create_histogram(merged_df, var, variables[[var]]))
combined_plots <- wrap_plots(plots, nrow = 2, ncol = 2)
combined_plots
```
## Descriptive Statistics by Outcome
```{r}
# Define variables
vars <- c("age", "gendera")
# Convert gendera to a factor
merged_df$gendera <- factor(merged_df$gendera, levels = c(1, 2), labels = c("Male", "Female"))
# Create and print Table 1
table1 <- CreateTableOne(vars = vars, strata = "outcome", data = merged_df)
print(table1, showAllLevels = TRUE)
```
# Data Transformation
## Create Age Groups
```{r}
# Create a new column for age groups based on age ranges.
merged_df$age_group <- cut(merged_df$age,
breaks = seq(0, 100, by = 10),
right = FALSE,
labels = c("0-9", "10-19", "20-29", "30-39", "40-49",
"50-59", "60-69", "70-79", "80-89", "90-99"))
```
## Recode Ethnicity
```{r}
# Create a mapping for ethnicity recoding.
ethnicity_mapping <- c(
"AMERICAN INDIAN/ALASKA NATIVE" = "American Indian/Alaska Native",
"ASIAN" = "Asian",
"ASIAN - ASIAN INDIAN" = "Asian",
"ASIAN - CAMBODIAN" = "Asian",
"ASIAN - CHINESE" = "Asian",
"ASIAN - FILIPINO" = "Asian",
"ASIAN - VIETNAMESE" = "Asian",
"BLACK/AFRICAN" = "Black/African",
"BLACK/AFRICAN AMERICAN" = "Black/African",
"BLACK/CAPE VERDEAN" = "Black/African",
"BLACK/HAITIAN" = "Black/African",
"HISPANIC OR LATINO" = "Hispanic/Latino",
"HISPANIC/LATINO - PUERTO RICAN" = "Hispanic/Latino",
"MIDDLE EASTERN" = "Middle Eastern",
"MULTI RACE ETHNICITY" = "Multi Race",
"OTHER" = "Other",
"PATIENT DECLINED TO ANSWER" = "Other",
"PORTUGUESE" = "Other",
"UNABLE TO OBTAIN" = "Other",
"UNKNOWN/NOT SPECIFIED" = "Other",
"WHITE" = "White",
"WHITE - OTHER EUROPEAN" = "White",
"WHITE - RUSSIAN" = "White"
)
# Recode the ethnicity column based on the mapping.
merged_df$main_ethnicity <- ethnicity_mapping[merged_df$ethnicity]
```
# Data Summarization
## Summarize by Ethnicity
```{r, fig.height=7}
# Calculate the counts and percentages for each ethnicity.
ethnicity_counts <- table(merged_df$main_ethnicity)
percentages <- prop.table(ethnicity_counts) * 100
# Create a dataframe with ethnicity and their corresponding percentages.
ethnicity_table <- data.frame(
Ethnicity = names(percentages),
Percentage = as.vector(percentages)
)
# Arrange the dataframe by percentage in descending order.
ethnicity_table <- ethnicity_table %>%
arrange(desc(Percentage))
# Reorder the factor levels of Ethnicity.
ethnicity_table$Ethnicity <- factor(ethnicity_table$Ethnicity, levels = ethnicity_table$Ethnicity)
# Create the bar plot with percentage labels for each ethnicity.
ggplot(ethnicity_table, aes(x = Ethnicity, y = Percentage, fill = Ethnicity)) +
geom_bar(stat = "identity") +
geom_text(aes(label = sprintf("%.1f%%", Percentage)),
vjust = -0.3, size = 3.5) +
labs(title = "Percentage of Each Ethnicity",
x = "Ethnicity",
y = "Percentage") +
theme_classic() +
theme(
axis.text.x = element_text(angle = 60, hjust = 1, size = 12),
axis.text.y = element_text(size = 12),
axis.title.x = element_text(size = 14),
axis.title.y = element_text(size = 14),
plot.title = element_text(size = 16, face = "bold")
)
```
## Summarize by Gender
```{r}
# Calculate the counts and percentages for each gender.
gender_counts <- table(merged_df$gender)
percentages <- prop.table(gender_counts) * 100
# Create a dataframe with gender and their corresponding percentages.
gender_table <- data.frame(
Gender = names(percentages),
Percentage = as.vector(percentages)
)
# Arrange the dataframe by percentage in descending order.
gender_table <- gender_table %>%
arrange(desc(Percentage))
# Reorder the factor levels of Gender.
gender_table$Gender <- factor(gender_table$Gender, levels = gender_table$Gender)
# Create the bar plot with percentage labels for each gender.
ggplot(gender_table, aes(x = Gender, y = Percentage, fill = Gender)) +
geom_bar(stat = "identity") +
geom_text(aes(label = sprintf("%.1f%%", Percentage)),
vjust = -0.3, size = 3.5) +
theme_minimal() +
labs(title = "Percentage of Each Gender",
x = "Gender",
y = "Percentage") +
theme_classic() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
```
## Summarize by Age Group
```{r}
# Calculate the counts and percentages for each age group.
age_group_counts <- table(merged_df$age_group)
percentages <- prop.table(age_group_counts) * 100
# Create a dataframe with age group and their corresponding percentages.
age_group_table <- data.frame(
AgeGroup = names(percentages),
Percentage = as.vector(percentages)
)
# Arrange the dataframe by percentage in descending order.
age_group_table <- age_group_table %>%
arrange(desc(Percentage))
# Reorder the factor levels of AgeGroup.
age_group_table$AgeGroup <- factor(age_group_table$AgeGroup, levels = age_group_table$AgeGroup)
# Create the bar plot with percentage labels for each age group.
ggplot(age_group_table, aes(x = AgeGroup, y = Percentage, fill = AgeGroup)) +
geom_bar(stat = "identity") +
geom_text(aes(label = sprintf("%.1f%%", Percentage)),
vjust = -0.3, size = 3.5) +
theme_minimal() +
labs(title = "Percentage of Each Age Group",
x = "Age Group",
y = "Percentage") +
theme_classic() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
```