heart_attack.Rmd

---
title: "Heart_Attack Analysis "
author: "Bernard Asante"
date: "2024-11-23"
output: html_document
---
##Loading library 
```{r, warning=FALSE}
library(tidyverse)
library(vtable)
```

##REading data 
```{r}
df <- read.csv('Heart Attack.csv')
df %>% 
  colnames()
```

##Data Exploration
```{r, warning=FALSE}
df %>% 
  glimpse()

df %>% 
  head()   #displaying first six rows

df %>% 
  tail()  #Displaying last 6 rows 
```

##Data cleaning 
```{r}
df <- df %>% 
  select(age, gender, impluse, pressurehight, pressurelow, glucose, troponin, class) %>% 
  rename(pulse = impluse, systolic = pressurehight, diastolic = pressurelow) %>% 
  mutate(`glucose_mmol/l` = glucose / 18) %>% 
  mutate(class= as.factor(class)) %>% 
  mutate(gender = recode(gender, 
                         `0` = 'Female',
                         `1` = 'Male')) %>% 
  select(-glucose)

head(df)


df %>% 
  summarise_all(~sum(is.na(.))) 

head(df)
```

##Data manipulation on categorical vaiables 

```{r}
df %>% 
  glimpse()   #observing columns with character or factor 

freq_gender <- df %>% 
  count(gender) %>% 
  rename(freq = n)

freq_gender

freq_class <- df %>% 
  count(class) %>% 
  rename(freq = n)
freq_class


# Contingency(Two by two )table
gender_class <- df %>% 
  select(gender, class) %>% 
  table()

gender_class
```




##Data manipulation on numeric variables 
```{r}
df_mean <- df %>% 
  summarise_if(is.numeric, mean)  #displaying mean fro all numeric values 

df_mean

mean_age <- df %>% 
  summarise(mean_age = mean(age))    #displaying mean for a single numeric value

mean_age



df_fivenum <- df %>% 
  summarise_if(is.numeric, fivenum)

st(df_fivenum)


```
## Data manipulation (cat and num variables)
```{r}
means_for_gender <- df %>% 
  group_by(gender) %>% 
  summarise_if(is.numeric, mean)

means_for_gender  

means_for_class <- df %>% 
  group_by(class) %>% 
  summarise_if(is.numeric, mean)

means_for_class # this can further be evaluated using t test 
```

##Data visualization 
#Visualizing categorical variables (count)
```{r}


bar_gender <- df %>% 
  ggplot(aes(gender, fill = gender))+
  geom_bar()+
  labs(title = 'Bar chart of gender distrubution')

bar_gender

bar_class <- df %>% 
  ggplot(aes(class, fill = class))+
  geom_bar()+
  labs(title = 'Bar chart of class distrubution')

bar_class

#Displaying across other cat variables 
bar_gender_class <- df %>% 
  ggplot(aes(gender, fill = gender))+
  geom_bar()+
  labs(title = 'Bar chart of gender distrubution betweeen class')+
  facet_grid(~class)

bar_gender_class


bar_class_gender <- df %>% 
  ggplot(aes(class, fill = class))+
  geom_bar()+
  labs(title = 'Bar chart of gender distrubution betweeen class')+
  facet_grid(~gender)

bar_class_gender
```
#Visualizing Numeric variables
```{r}
hist_age <- df %>% 
  ggplot(aes(age, fill = gender))+
  geom_histogram(bin = 10)+
  labs(title = 'Histogram for age')
hist_age

box_age <- df %>% 
  ggplot(aes(gender, age, fill = gender))+
  geom_boxplot()+
  facet_grid(~class)+
  labs(title = 'Boxplot for age bewteen class')
box_age


hist_troponin <- df %>% 
  ggplot(aes(troponin, fill = class))+
  geom_histogram(bin = 20)+
  labs(title = 'Histogram for tropnin ')+
  facet_grid(~class)
hist_troponin

box_troponin <- df %>% 
  ggplot(aes(gender, troponin, fill = gender))+
  geom_boxplot()+
  facet_grid(~class)+
  labs(title = 'Boxplot for troponin bewteen class')
box_troponin


#Scatter plot 
glucose_troponin_point <- df %>% 
  ggplot(aes(`glucose_mmol/l`, troponin, colour = class))+
  geom_point()+
  labs(title = 'Scatterplot for glucose and troponin')

glucose_troponin_point

age_troponin_point <- df %>% 
  ggplot(aes(age, troponin, colour = class))+
  geom_point()+
  geom_smooth()+
  labs(title = 'Scatterplot for age and troponin')

age_troponin_point
```
##Testing hypothesis 
#Chi_square
```{r}
chi_gender_class <- df %>% 
  select(gender,class) %>% 
  table()
chi_gender_class  #Same as gender_class table 

chisq.test(chi_gender_class)
```
# t test 
```{r}
test_gender_troponin <- df %>% 
  select(gender, troponin) %>% 
  t.test(troponin~gender, data = .)
test_gender_troponin

test_gender_age <- df %>% 
  select(gender, age) %>% 
  t.test(age~gender, data = .)
test_gender_age
```
#Correlation 
```{r}
cor.test(df$age, df$troponin)
```
#Regression models 
```{r, warning=FALSE}
model <- glm(class ~ age + gender + troponin, 
             data = df, 
             family = binomial)
summary(model)

```
#Visualizing model 
```{r}
library(ggplot2)

# Extract coefficients
coefficients <- summary(model)$coefficients
coeff_df <- data.frame(
  Predictor = rownames(coefficients),
  Estimate = coefficients[, "Estimate"],
  StdError = coefficients[, "Std. Error"],
  pValue = coefficients[, "Pr(>|z|)"]
)

# Plot coefficients
ggplot(coeff_df, aes(x = Predictor, y = Estimate)) +
  geom_point() +
  geom_errorbar(aes(ymin = Estimate - StdError, ymax = Estimate + StdError), width = 0.2) +
  theme_minimal() +
  labs(title = "Logistic Regression Coefficients", y = "Estimate", x = "Predictor")
```
#Linear model 
```{r}
linear_model <- lm(troponin ~ age+gender+class+`glucose_mmol/l`, data = df)
summary(linear_model)
```