Predictive Maintenance Data Visualization

Table of Contents

• Introduction
• Data Understanding
• Business Understanding
• Data Explanation
• Visualizations
• Running the Visualizations
• Contributors
• Acknowledgments

Introduction

In today's competitive industrial landscape, efficiency and reliability in production are paramount. This project aims to merge advanced data analytics with traditional manufacturing to enhance operational excellence. Utilizing detailed datasets from industrial environments, the goal is to create a predictive maintenance framework to foresee equipment failures and inefficiencies.

Understanding Data

This section delves into the significance of each data entry, capturing the dynamic nature of industrial processes. Here, we explore the stories behind various parameters like air and process temperatures, rotational speeds, torque, and tool wear, and transform these into actionable insights.

Understanding Business

The project extends beyond theoretical models to provide practical, impactful solutions in the industrial sector. Here, the focus is on aligning predictive insights with the strategic goals of businesses, translating into measurable benefits like cost savings and improved productivity.

Data Explanation

This part details the various data types used in the project, including:

• Identifier Data: UDI (Unique Identifier), Product ID.
• Categorical Data: Machine Types.
• Continuous Numerical Data: Air temperature, Process temperature, Rotational speed, Torque, Tool wear.
• Binary Data: Machine failure indicators.

Visualizations

Key visualizations include:

1. Correlations of Industrial Machine Parameters with Failure Events
2. Density Distributions of Machine Parameters by Type
3. Interactive Analysis of Process Temperature Variability by Machine Failure Type
4. Interactive Process Temperature vs. Power Output Colored by Error Type
5. Machine Type Frequency and Failure Incidence
6. Probability of Machine Failure Across Various Operational Parameters

Running the Visualizations

To generate any of these figures, run the main_code script and select the desired figure from the options provided. This script is designed to facilitate the visualization of the analyzed data effectively.

Contributors

• Yogev Attias
• Shir Michal Grief