Competition - Predicting industrial machine downtime 🔧 - Level 2

# Full Machine Downtime Analysis with Predictive Modeling

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import classification_report, confusion_matrix

# Set style
sns.set_style("whitegrid")
plt.rcParams['figure.figsize'] = (10, 6)
plt.rcParams['font.family'] = 'sans-serif'
plt.rcParams['font.sans-serif'] = ['Inter']

# --- Load Data ---
df = pd.read_csv('machine_downtime.csv')
df['Date'] = pd.to_datetime(df['Date'])
df['Downtime'] = df['Downtime'].apply(lambda x: 1 if x == 'Machine_Failure' else 0)

# --- Correlation Analysis ---
columns_to_exclude = ['Downtime', 'Assembly_Line_No']
numerical_cols = df.select_dtypes(include=np.number).columns
operational_cols = [col for col in numerical_cols if col not in columns_to_exclude]

# Calculate and plot correlation matrix
correlation_matrix = df[operational_cols].corr()
plt.figure(figsize=(12, 10))
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', fmt=".2f", linewidths=.5, cbar_kws={'label': 'Correlation Coefficient'})
plt.title('Correlation Matrix of Operational Parameters', fontsize=16)
plt.xticks(rotation=45, ha='right')
plt.yticks(rotation=0)
plt.tight_layout()
plt.show()

# --- Downtime Over Time ---
downtime_daily = df[df['Downtime'] == 1].groupby(['Date', 'Machine_ID']).size().unstack(fill_value=0)
plt.figure(figsize=(15, 7))
for machine_id in downtime_daily.columns:
    plt.plot(downtime_daily.index, downtime_daily[machine_id], marker='o', linestyle='-', label=f'Machine {machine_id}')
plt.title('Daily Downtime Events per Machine')
plt.xlabel('Date')
plt.ylabel('Downtime Events')
plt.legend()
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

# --- Boxplots of Key Features vs Downtime ---
key_features = [
    'Hydraulic_Oil_Temperature', 'Spindle_Vibration', 'Tool_Vibration', 'Torque(Nm)',
    'Coolant_Temperature', 'Coolant_Pressure(bar)', 'Air_System_Pressure(bar)',
    'Voltage(volts)', 'Spindle_Speed(RPM)', 'Cutting_Force(KN)'
]
plt.figure(figsize=(18, 15))
for i, feature in enumerate(key_features):
    if feature in df.columns:
        plt.subplot(4, 3, i + 1)
        sns.boxplot(x='Downtime', y=feature, data=df, palette='viridis')
        plt.title(f'{feature} vs. Downtime')
        plt.xlabel('Downtime (0: No, 1: Yes)')
        plt.ylabel(feature)
        plt.xticks([0, 1], ['Operational', 'Downtime'])
plt.tight_layout()
plt.show()

# --- Predictive Modeling ---
X = df[operational_cols]
y = df['Downtime']
X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y, test_size=0.2, random_state=42)

model = RandomForestClassifier(random_state=42)
model.fit(X_train, y_train)
y_pred = model.predict(X_test)

# --- Evaluation ---
print("--- Classification Report ---")
print(classification_report(y_test, y_pred))
print("--- Confusion Matrix ---")
print(confusion_matrix(y_test, y_pred))

# --- Feature Importance ---
importances = model.feature_importances_
feat_imp = pd.Series(importances, index=X.columns).sort_values(ascending=False)
plt.figure(figsize=(10, 6))
sns.barplot(x=feat_imp.values, y=feat_imp.index, palette='crest')
plt.title('Feature Importance from Random Forest Model')
plt.xlabel('Importance')
plt.ylabel('Features')
plt.tight_layout()
plt.show()

# --- Recommendations ---
print("\n--- Recommendations ---")
print("1. Develop a predictive maintenance system using the trained model.")
print("2. Monitor critical features such as Hydraulic_Oil_Temperature and Vibration levels.")
print("3. Set dynamic thresholds based on historical values and model outputs.")
print("4. Prioritize monitoring and maintenance of Machine 'M-03' due to frequent downtime.")
print("5. Continuously retrain the model using updated sensor data to maintain accuracy.")