🧠 Daily Study Log [2025-07-18]
Today was a diverse study day across ideation, coding practice, paper review, and spatial data exploration.
I’m also considering transitioning this blog into a more polished portfolio over time.

🧠 Project: Seoul Youth Footprint — Mapping Young Mobility for Smarter Urban Policy

This project explores where young people in Seoul actually spend their time using vehicle ownership stats and optionally GPS-based voluntary tracking.
It aims to support location-aware youth programs by revealing movement patterns that go beyond traditional demographic targeting.

Problem:
Policy decisions often rely on static population data, ignoring behavioral patterns like actual destinations and emerging urban spaces popular with youth.

Approach:

• Gather youth vehicle data and demographics from public portals
• Tag Points of Interest (POIs) and apply clustering
• Optionally include consent-based GPS samples for richer movement profiles
• Visualize on maps to guide youth-centered programs and infrastructure planning

🗺 GIS Project: Seoul Smoking Zone Mapping

This project visualizes designated public smoking zones in Seoul, starting with Yongsan-gu.
It uses open data and geocoding to plot real-world coordinates and supports policy transparency.

Scope:
Only districts with usable coordinate-level data are included (Yongsan-gu for now). Other areas will be added as data quality permits.

Purpose:

• Inform public health planning
• Analyze spatial clustering of legal zones
• Serve as a base for potential zone recommendations or impact studies

📄 Paper Review — Unsupervised 3D Pose Estimation for Hierarchical Dance Video Recognition

Focus: Section 2 — Computational Approach
Link: https://arxiv.org/abs/2109.09166
Repo: github.com/hojjang98/Paper-Review

Reflection:
Today’s section introduced a 3-stage dancer tracking algorithm that handles occlusion using motion prediction and histogram matching.
The approach is elegant and practical, especially in crowded multi-person dance scenarios.
Next up: the adversarial 3D pose lifting part, which seems more technical and will require closer attention.

🧠 SQL Practice — Query Prediction & Debugging

What I Did:
Practiced predicting SQL query results before execution and focused on error tracing without relying on outputs. Worked with JOINs, GROUP BY, HAVING, and subqueries.

Insights:

• Training yourself to guess the output sharpens schema understanding
• Misuse of HAVING vs WHERE is still a recurring bug
• Will continue refining logic-driven query thinking

✅ TL;DR

📍 Project: Designed a youth mobility analysis project using behavioral + vehicle data
📍 GIS: Started mapping legal smoking zones — currently Yongsan-gu only
📍 Paper: Learned occlusion-aware tracking in dance recognition pipeline
📍 SQL: Focused on query logic without running them — great debugging exercise

💡 Note: I’m considering turning this blog into a more portfolio-style archive of projects, not just logs. The formatting and content depth will gradually reflect that shift.