About the Platform

About Data Driven Models

A research data platform for AI-driven quality control in manufacturing.

Data Driven Models is a web-based platform developed at Blekinge Institute of Technology that brings together research data management, machine learning deployment, and cross-institutional collaboration — enabling research teams and industry partners to turn simulation data into actionable quality predictions through a secure, role-based interface.

Capabilities

Platform Capabilities

Everything you need for data-driven manufacturing research.

📂
Research Data Management

Hierarchical data organisation by project, material grade, and data type with filesystem-backed storage.

🤖
ML Model Deployment

Version-controlled model library supporting ANN, Random Forest, and XGBoost with per-part tracking.

🎯
Interactive Applications

Real-time quality prediction tools with SHAP explainability and process optimisation.

💬
AI-Powered Analysis

Built-in AI assistant for querying data, generating charts, and exploring model strategies.

👥
Multi-Project Collaboration

Role-based access control (Admin, Researcher, Viewer) across multiple research projects.

🔍
Data Explorer

Browse and download datasets by category, material grade, or data type.

📑
Research Publications

Centralised access to conference papers, journal articles, and demo videos.

🚦
Quality Traffic Light

GREEN / YELLOW / RED status system for instant quality assessment on every prediction.

Technology

Technology Stack

Built on proven, open-source technologies.

🐍
Python

Core language for backend and ML pipelines

🌐
Flask

Lightweight web framework with Jinja2 templates

🧠
scikit-learn

MLP neural networks, Random Forest, and XGBoost models

📊
Chart.js

Interactive data visualisation and dashboards

🗄
SQLite / PostgreSQL

Relational database for structured research data

🐳
Docker

Containerised deployment with Docker Compose

🚀
Nginx

Reverse proxy, SSL termination, and static serving

🔬
SHAP

Explainable AI for transparent model predictions

Collaborate

Interested in Collaboration?

We welcome research partners, industry collaborators, and academic institutions interested in data-driven manufacturing and circular steel innovation.

Get in Touch View Research