Automated Quality Assurance Dashboard for ChargePoint

Introduction

As part of ENGG 199 - Special Topics in Engineering Sciences, I worked on a full-stack development project focused on improving manufacturing quality assurance for ChargePoint. This course provided an opportunity to apply software development, cloud infrastructure, and computer vision techniques in a real-world setting. The goal was to automate defect detection for EV chargers using a React-based dashboard and AWS services.

Problem

Manufacturing high-quality EV chargers requires rigorous quality control, but the existing process relied heavily on manual inspections, leading to delays, incomplete data, and inefficiencies. ChargePoint needed an automated system to capture defect data in real time, reduce inspection time per unit, and improve traceability for defect analysis.

Solution

I built a React-powered dashboard that integrates computer vision, cloud computing, and real-time analytics to monitor key production metrics, including first pass yield, retest and rework rates, final yield, cycle time, and takt time. The dashboard allows users to search and filter quality control data by serial number, factory location, and pass or fail status, providing engineers with instant access to critical insights.

Screenshot of the dashboard I developed and presented for ChargePoint. Built with React, AWS Amplify, and GraphQL to automate quality assurance and defect tracking before the rise of GenAI.

Recent capture data and detailed charger view from the dashboard I built for ChargePoint.

Cloud Infrastructure and Backend

Originally, the project was set up with MongoDB, but I transitioned to AWS Amplify and Cognito for a scalable authentication and data management system. Rather than building a traditional backend with custom server logic, I leveraged AWS’s managed services to handle authentication, database interactions, and API management. I configured AWS AppSync for efficient GraphQL querying and implemented AWS S3 for secure storage and retrieval of inspection images. This cloud-based setup ensured real-time data accessibility and automation while minimizing the need for direct backend maintenance.

To automate quality checks, the system captured barcode and component images, analyzing them with SIFT and other image processing algorithms in Python. This enabled automated defect detection, reducing the need for manual inspection and improving error traceability.

Error Tracking and Visualization

I developed an interactive dashboard that categorized and visualized manufacturing defects. Using Recharts.js, I built bar charts for error frequency analysis, helping engineers identify recurring issues. Additionally, I implemented moment.js to calculate time-based metrics such as units per hour and cycle time, providing deeper insights into production efficiency.

Statistical Analysis and Performance Metrics

The dashboard calculated first pass yield as the ratio of chargers that passed inspection on the first attempt to the total number of chargers produced. This was updated in real time and displayed as a percentage.

The retest rate was determined by the proportion of chargers flagged for rework after failing the initial inspection. The system also calculated final yield, which measured the percentage of chargers that ultimately passed inspection, including those that required rework.

Cycle time was computed by measuring the time difference between the first recorded inspection of the day and the last, divided by the number of chargers processed. The system also estimated units per hour using timestamp data from inspection logs, calculating the rate at which chargers were passing through quality control.

Impact

This project merged hardware, software, and cloud technologies, improving production efficiency, defect traceability, and real-time quality monitoring. By automating quality control processes, the system reduced inspection time per unit and provided engineers with actionable insights to improve manufacturing performance.

Due to NDA, I can discuss technical details upon request.