Driving Behaviour EDA and Classification

!pip install kaggle

import json
import os
from kaggle.api.kaggle_api_extended import KaggleApi

# Path to where the Kaggle API key is stored
kaggle_json_path = os.path.join(os.path.expanduser('~'), '.kaggle', 'kaggle.json')

# Load the Kaggle API key
with open(kaggle_json_path) as f:
    kaggle_api_key = json.load(f)

# Initialize Kaggle API
api = KaggleApi()

# Set Kaggle API credentials directly
api.set_config_value('username', 'rrivas2')
api.set_config_value('key', '0312095f24321eb24b9dce6f85c54b5b')

# Download the dataset
api.dataset_download_files(dataset='outofskills/driving-behavior', unzip=True)

import pandas as pd
train = pd.read_csv('train_motion_data.csv')

train.head()

train.describe()

train.info()

train.shape

import matplotlib.pyplot as plt

# Filtering the dataset for the 'slowdriving' category
slowdriving_data = train[train['Class'] == 'SLOW']

# Plotting acceleration data (AccX, AccY) for the 'slowdriving' category
plt.figure(figsize=(15, 6))
plt.plot(slowdriving_data['Timestamp'], slowdriving_data['AccX'], label='AccX')
plt.plot(slowdriving_data['Timestamp'], slowdriving_data['AccY'], label='AccY')

plt.xlabel('Timestamp')
plt.ylabel('Acceleration')
plt.title('Acceleration Data (AccX, AccY) Over Time for Slow Driving')
plt.legend()
plt.show()

import matplotlib.pyplot as plt

# Filtering the dataset for the 'slowdriving' category
normaldriving_data = train[train['Class'] == 'NORMAL']

# Plotting acceleration data (AccX, AccY) for the 'slowdriving' category
plt.figure(figsize=(15, 6))
plt.plot(normaldriving_data['Timestamp'], normaldriving_data['AccX'], label='AccX')
plt.plot(normaldriving_data['Timestamp'], normaldriving_data['AccY'], label='AccY')

plt.xlabel('Timestamp')
plt.ylabel('Acceleration')
plt.title('Acceleration Data (AccX, AccY) Over Time for Normal Driving')
plt.legend()
plt.show()

import matplotlib.pyplot as plt

# Filtering the dataset for the 'slowdriving' category
aggressivedriving_data = train[train['Class'] == 'AGGRESSIVE']

# Plotting acceleration data (AccX, AccY) for the 'slowdriving' category
plt.figure(figsize=(15, 6))
plt.plot(aggressivedriving_data['Timestamp'], aggressivedriving_data['AccX'], label='AccX')
plt.plot(aggressivedriving_data['Timestamp'], aggressivedriving_data['AccY'], label='AccY')

plt.xlabel('Timestamp')
plt.ylabel('Acceleration')
plt.title('Acceleration Data (AccX, AccY) Over Time for Aggressive Driving')
plt.legend()
plt.show()

import matplotlib.pyplot as plt

# Plotting gyro data (GyroY, GyroZ) for the 'slowdriving' category
plt.figure(figsize=(15, 6))
plt.plot(slowdriving_data['Timestamp'], slowdriving_data['GyroY'], label='GyroY')
plt.plot(slowdriving_data['Timestamp'], slowdriving_data['GyroZ'], label='GyroZ')

plt.xlabel('Timestamp')
plt.ylabel('Gyro readings')
plt.title('Gyro Data (GyroY, GyroZ) Over Time for Slow Driving')
plt.legend()
plt.show()

import matplotlib.pyplot as plt

# Plotting gyro data (GyroY, GyroZ) for the 'normaldriving' category
plt.figure(figsize=(15, 6))
plt.plot(normaldriving_data['Timestamp'], normaldriving_data['GyroY'], label='GyroY')
plt.plot(normaldriving_data['Timestamp'], normaldriving_data['GyroZ'], label='GyroZ')

plt.xlabel('Timestamp')
plt.ylabel('Gyro readings')
plt.title('Gyro Data (GyroY, GyroZ) Over Time for Normal Driving')
plt.legend()

import matplotlib.pyplot as plt

# Plotting gyro data (GyroY, GyroZ) for the 'aggressivedriving' category
plt.figure(figsize=(15, 6))
plt.plot(aggressivedriving_data['Timestamp'], aggressivedriving_data['GyroY'], label='GyroY')
plt.plot(aggressivedriving_data['Timestamp'], aggressivedriving_data['GyroZ'], label='GyroZ')

plt.xlabel('Timestamp')
plt.ylabel('Gyro readings')
plt.title('Gyro Data (GyroY, GyroZ) Over Time for Aggressive Driving')
plt.legend()
plt.show()

# Assume df is your DataFrame containing the dataset
class_mapping = {
    'SLOW': 0,
    'NORMAL': 1,
    'AGGRESSIVE': 2
}

# Add a new column with numerical labels
train['Class_Labels'] = train['Class'].map(class_mapping)

# Print the mapping of original class labels to numerical labels
print("Class Label Mapping:")
for label, numeric_label in class_mapping.items():
    print(f"{label}: {numeric_label}")

y_train = train['Class_Labels']
X_train = train.drop(columns=['Class','Class_Labels','Timestamp'])
X_train.head()