skip to content
文章 · 2026年6月 Posts · June 2026

Field Day

作者 By Yc Lin·2026年6月25日25 June 2026·12 min read

UW–Madison Field Day Lab — Real-Time Educational Game Analytics Platform

1. Overview

Worked as a Data Science Product Intern at UW–Madison Field Day Lab to design and deliver a real-time analytics platform for educational games. I owned product requirements, dashboard design, and technical implementation of core visualization features.

2. STAR Breakdown

Situation

Researchers studying educational games relied primarily on offline log analysis, making it difficult to observe player behavior during live gameplay. Existing analytics tools were fragmented across games and lacked a unified real-time monitoring interface.

Task

Design a product that enables researchers to monitor gameplay events in real time, analyze player behavior across multiple educational games, and provide a scalable analytics experience without disrupting the existing logging infrastructure.

Action

  • Conducted discussions with researchers to understand monitoring workflows and translated research needs into product requirements.

  • Defined the dashboard information architecture, including Session-, Player-, and Population-level analytics, game selection, event monitoring, and multi-dimensional filtering.

  • Evaluated multiple technical approaches for real-time event delivery, balancing implementation risk with compatibility of the existing PHP logging system, and adopted a Flask + Flask-SocketIO relay instead of replacing the production logger.

  • Developed D3.js visualization components and responsive dashboard interfaces to present live gameplay events and player metrics.

Result

  • Delivered a real-time analytics dashboard capable of supporting educational game event monitoring.

  • Established a unified analytics workflow across more than five educational games.

  • Successfully connected the existing logging pipeline with a real-time visualization interface without requiring major changes to the production logging system.

3. Product Thinking Highlights

  • Identified researchers—not players—as the primary users and designed the dashboard around their observation workflow.

  • Prioritized compatibility with the existing production logger over rebuilding the data pipeline from scratch.

  • Structured analytics into Session, Player, and Population levels to support different research questions.

  • Combined product requirement definition with technical solution design to bridge research needs and engineering implementation.

4. Challenges & How I Solved Them

Challenge 1: Existing logging infrastructure was already in production.

Replacing the PHP logger would introduce unnecessary risk and potentially affect ongoing research.

Solution: Designed a relay architecture using Flask and WebSocket technology to stream events in real time while preserving the existing logging pipeline.

Challenge 2: Requirements evolved during development.

Researchers refined their analytics needs as they explored prototype dashboards.

Solution: Designed modular dashboard components and filtering capabilities so new metrics and visualizations could be added without redesigning the entire interface.

Challenge 3: Multiple educational games generated different event streams.

Different games produced heterogeneous event data, making unified analysis difficult.

Solution: Designed a common analytics interface with game selection and Session-, Player-, and Population-level views that worked across multiple games.

5. Impact Summary

  • Designed the product requirements for a real-time educational game analytics dashboard.

  • Delivered live event visualization using D3.js and WebSocket-based data streaming.

  • Built a unified analytics experience supporting multiple educational games.

  • Bridged user requirements, product design, and technical implementation within a research environment.

6. Reflection

This project changed how I think about product management for technical platforms. Rather than focusing only on adding features, I learned that understanding existing systems and choosing solutions with lower implementation risk can create greater long-term value. Designing around researcher workflows instead of technical possibilities also reinforced the importance of starting with user needs before making engineering decisions.

评论Comments