Summary

SHU Prophet is an academic-driven platform that productizes cutting-edge time series forecasting research into an intelligent decision-making tool. It targets researchers, enterprise decision-makers, and data analysts, offering an interactive, AI-powered web experience with advanced, research-backed predictive models and conversational analysis capabilities. The platform aims to democratize sophisticated time series analysis beyond traditional research settings.

How It Works

The platform integrates six novel time series forecasting models derived from CCF-rated academic publications, employing techniques like wavelet transforms, scattering transforms, and diffusion models. It features a conversational AI assistant for natural language data analysis, a dual-engine prediction system (ARIMA + custom multi-agent), and Chain-of-Thought reasoning for transparent, in-depth insights. This approach combines state-of-the-art academic innovation with user-friendly interaction.

Quick Start & Requirements

Deployment is streamlined via Zeabur one-click setup or Docker, requiring only an OpenAI-compatible API key (e.g., Moonshot, Zhipu GLM) and PostgreSQL for production. Local development involves standard backend (Flask) and frontend (Vue/Vite) setup. The platform is browser-accessible, eliminating environment configuration for end-users.

Highlighted Details

• Six proprietary time series forecasting models based on CCF-B/C conference papers (ICASSP, ICANN, ICIC).
• Conversational AI assistant enabling no-code data analysis, trend identification, anomaly detection, and automated report generation.
• Dual-engine prediction combining ARIMA with a self-developed multi-Agent collaborative intelligent engine.
• Interactive web interface featuring multi-model visualization, standardized metric evaluation, and community knowledge sharing.

Maintenance & Community

Developed by an undergraduate team from Shanghai University's Computer Engineering and Science School. Primary contact is Wei Li. Includes community features like a sharing square and a user points/level system, suggesting user engagement is a focus, though long-term maintenance signals are not detailed.

Licensing & Compatibility

Licensed under the Apache License 2.0, permitting broad commercial use and integration into closed-source projects.

Limitations & Caveats

The platform relies on external LLM API keys, incurring potential costs and usage limits managed via a points system. As an undergraduate project, long-term maintenance and support maturity may require further evaluation.