Brief description of the data set

This Health Care: Heart attack dataset is a part of a big database that originally contains 76 attributes. This subset is created by using only 14 attributes such as age, sex, chest pain type, etc. of the patients. This dataset is highly popular for the prediction of heart attack for ML researches. This dataset contains 14 columns in total. The heart disease in the patient is given by the "target" field and the rest of the 13 columns are the features column. The total sample size of this dataset is 303. All feature columns are integer types except the "oldpeak" which is specified as a float.

age: depict how old the person is.
sex: shows the gender of a person.
chest pain type(cp): The patient's chest pain is categorized according to 4 different values from 0 to 3.
resting blood pressure(trestbps): Resulting in heart-pumping, circulation of blood leads to pressure against the walls of blood vessels. For a healthy person, normal resting blood pressure is 120 mm/Hg.
serum cholesterol (chol): It is closely related to heart disease because the low level of lipoproteins can cause a restriction or blockage in the blood flow. 125-200 mg/dL is considered normal for a healthy person.
fasting blood sugar (fbs): It shows the blood sugar level of a person by taking a blood sample after an overnight fast. This feature is given whether a patient's blood sugar level is higher than 120 mg/dL or not.
resting electrocardiographic results (restecg): Resting electrocardiography (ECG) is used for detecting abnormalities such as arrhythmias and evidence of coronary heart disease. Values are shown 0 to 2.
maximum heart rate achieved (thalach): It depicts the person's maximum heartbeat. If disease weakens the heart, sufficient blood can not reach the organs.
exercise induced angina (exang): Angina is caused by low blood flow to the heart muscle. Ischemia is the result of this reduced blood flow to the heart muscle. This feature shows whether a person has angina or not.
old peak (ST depression induced by exercise) (oldpeak): It can be used for the detection of coronary atherosclerosis disease by using the ECG. It can be also related to coronary artery disease.
slope (the slope of the peak exercise ST): It shows the slope of the ST segment caused by the exercise-induced increments in heart rate.
number of major vessels (ca): Common arterial diseases lead to the narrowing of the blood vessels. The low level of oxygen and nutrition required for the heart to operate properly can not be achieved.
defect level (thal): This feature contains 4 levels from 0 to 3. 0 means a normal defect, 1 means a fixed defect and 2 is a reversible defect. There is also a level 3 in the feature column but they do not provide information for this value.
target: This is the target column of the dataset. Feature columns are used for the prediction of it. 0 means less chance of heart attack and 1 means more chance of heart attack.

Main objective: Evaluation of whether a person will have a heart attack or not plays an important role for many patients who have a high risk of coronary artery disease risk. High accuracy of prediction can help us to save many patients by taking protective precautions. Determination of a standardized set of features to assess the heart attack is a vital process and human intervention leads to unstable decisions. Because of these reasons, the main consideration of this project is the prediction of the heart attack for a patient according to the given features by using Machine Learning Classification algorithms.

Plans for Data Exploration

Data exploration will start with analyzing the categorical columns of the dataset. Basic actions such as counting unique values, calculating their frequencies in the distribution will be examined. Taking these actions will show what data is mainly about and the content of the categorical features.