Data Lab Manual

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

# Load the dataset
df = pd.read_csv('Dataset - conditionMonitoring.csv')

# Basic exploration
print(f"Dataset Shape: {df.shape}")
print(f"\nColumn Names:\n{df.columns.tolist()}")
print(f"\nData Types:\n{df.dtypes}")
print(f"\nFirst 5 Rows:\n{df.head()}")
print(f"\nBasic Statistics:\n{df.describe()}")

# Separate columns by axis
x_cols = [col for col in df.columns if col.startswith('xAcc')]
y_cols = [col for col in df.columns if col.startswith('yAcc')]
z_cols = [col for col in df.columns if col.startswith('zAcc')]

print(f"X-axis columns: {len(x_cols)}")
print(f"Y-axis columns: {len(y_cols)}")
print(f"Z-axis columns: {len(z_cols)}")

# Extract frequencies from column names
frequencies = [int(col[4:7]) for col in x_cols]
print(f"\nFrequencies (Hz): {frequencies}")

# Examine the Condition column
print(f"\nCondition values:\n{df['Condition'].value_counts()}")

# Plot average vibration spectrum for each axis
fig, axes = plt.subplots(3, 1, figsize=(12, 10))

# Calculate mean values for each frequency
x_means = df[x_cols].mean().values
y_means = df[y_cols].mean().values
z_means = df[z_cols].mean().values

# X-axis spectrum
axes[0].bar(frequencies, x_means, color='blue', alpha=0.7)
axes[0].set_title('X-Axis Average Vibration Spectrum')
axes[0].set_xlabel('Frequency (Hz)')
axes[0].set_ylabel('Acceleration')

# Y-axis spectrum
axes[1].bar(frequencies, y_means, color='green', alpha=0.7)
axes[1].set_title('Y-Axis Average Vibration Spectrum')
axes[1].set_xlabel('Frequency (Hz)')
axes[1].set_ylabel('Acceleration')

# Z-axis spectrum
axes[2].bar(frequencies, z_means, color='red', alpha=0.7)
axes[2].set_title('Z-Axis Average Vibration Spectrum')
axes[2].set_xlabel('Frequency (Hz)')
axes[2].set_ylabel('Acceleration')

plt.tight_layout()
plt.show()

# Calculate statistical features for each sample
def extract_vibration_features(row, axis_cols):
    values = row[axis_cols].values
    return {
        'mean': np.mean(values),
        'std': np.std(values),
        'max': np.max(values),
        'min': np.min(values),
        'range': np.max(values) - np.min(values),
        'rms': np.sqrt(np.mean(values**2))
    }

# Extract features for X-axis
x_features = df.apply(lambda row: extract_vibration_features(row, x_cols), axis=1)
x_features_df = pd.DataFrame(x_features.tolist())
x_features_df.columns = ['x_' + col for col in x_features_df.columns]

print("X-axis Feature Statistics:")
print(x_features_df.describe())

from scipy import stats

# Calculate overall vibration energy (RMS across all frequencies)
df['total_rms'] = np.sqrt((df[x_cols]**2).mean(axis=1) +
                          (df[y_cols]**2).mean(axis=1) +
                          (df[z_cols]**2).mean(axis=1))

# Calculate Z-scores
df['rms_zscore'] = stats.zscore(df['total_rms'])

# Identify potential anomalies (|Z| > 2)
anomalies = df[np.abs(df['rms_zscore']) > 2]
print(f"Number of potential anomalies: {len(anomalies)}")
print(f"Anomaly percentage: {len(anomalies)/len(df)*100:.2f}%")