README.md

---
license: apache-2.0
---
Predictive Maintenance Engine Dataset
Dataset Overview

This dataset contains engine sensor readings collected for predictive maintenance analysis. The objective is to develop machine learning models capable of identifying whether an engine is operating normally or requires maintenance intervention based on operational and thermal sensor measurements.

The dataset is designed to support predictive maintenance applications for various engine-driven systems including automobiles, portable generators, lawnmowers, and compact industrial machinery.

The sensor values represent realistic operational behavior across both small and large engine environments.

Business Problem

Unexpected engine failures can lead to:

Expensive repairs
Operational downtime
Reduced equipment lifespan
Safety risks
Productivity losses for fleet operators and manufacturers

Traditional maintenance strategies are often reactive or based on fixed schedules, which may either:

miss early warning signs of failure, or
lead to unnecessary servicing costs.

This dataset enables the development of machine learning solutions that support proactive maintenance scheduling using engine telemetry and sensor analytics.

Objective

The primary objective is to predict engine condition using sensor data and classify whether an engine:

is operating normally, or
requires maintenance attention.

The dataset supports supervised machine learning classification tasks for predictive maintenance systems.

Dataset Features
Feature Name	Description
Engine rpm	Engine rotational speed measured in revolutions per minute (RPM)
Lub oil pressure	Lubricating oil pressure responsible for reducing engine friction
Fuel pressure	Fuel delivery pressure influencing combustion efficiency
Coolant pressure	Cooling system pressure used for thermal regulation
lub oil temp	Lubricating oil temperature affecting lubrication quality
Coolant temp	Engine coolant temperature used to monitor overheating conditions
Engine Condition	Target variable representing engine health condition (0 = Normal, 1 = Maintenance Required/Faulty)
Target Variable

The target variable is:

Engine Condition

Classification labels:

0 → Normal engine operation
1 → Engine requires maintenance / faulty condition
Intended Use

This dataset is intended for:

Predictive maintenance modeling
Binary classification tasks
Sensor analytics research
Machine learning experimentation
Automotive maintenance optimization
Fleet management analytics

Possible algorithms include:

Decision Trees
Random Forest
Gradient Boosting
AdaBoost
XGBoost
Deep Learning models
Machine Learning Applications

Potential applications include:

Real-time engine monitoring systems
Maintenance scheduling optimization
Fleet reliability analysis
Failure risk prediction
Intelligent maintenance alerts
Data Characteristics
Structured tabular dataset
Numerical sensor readings
Binary classification target
Suitable for supervised learning
Appropriate for ensemble learning techniques
Expected Insights

The dataset can help identify:

overheating patterns,
lubrication-related anomalies,
pressure instability,
abnormal operational conditions,
and sensor relationships associated with engine degradation.
Limitations
The dataset does not include timestamp-based sequential behavior.
External environmental variables are not included.
Failure severity levels beyond binary classification are not provided.
Ethical Considerations

This dataset does not contain:

personally identifiable information,
user-sensitive data,
or confidential operational records.

The dataset is intended purely for educational, research, and predictive maintenance modeling purposes.

Citation

If using this dataset for academic or educational purposes, please cite the corresponding project repository or submission.

Maintainer

Dataset maintained as part of a Predictive Maintenance Machine Learning Capstone Project.