"Elegance is when precision meets purpose." — My mantra for both golf and code.
Hi, I’m Lydiah—a systems thinker, tech lover, and woman on a mission to build intelligent tools that make elite performance accessible to anyone.
This year, I’m launching a bold new project called SwingSense, where I’ll apply my skills in machine learning, computer vision, and data science to one of the most elegant yet data-rich sports on earth: golf.
In this post (the first in a build-in-public series), I’ll introduce you to the concept, motivation, and the tech backbone powering this portfolio project. And trust me—it’s not just about swings and stats. It’s about changing how we learn mastery, forever.
🎯 The Vision: Golf Coaching Reinvented with AI
Golf is a game of inches, intuition, and iteration. Yet for most aspiring golfers, quality coaching is expensive, subjective, and geographically limited.
What if I could change that?
SwingSense is my attempt to:
- Analyze golf swings through pose estimation and kinematic data
- Compare those movements to elite professionals
- Offer real-time, explainable, and personalized feedback
- Track a user’s improvement over time using machine learning
We’re not just building a tool. We’re architecting an experience—where feedback feels like a personal caddie whispering insights right in your ear.
🧠 The Tech Stack: Marrying AI with Sports Science
To bring this vision to life, I’m engineering an end-to-end system with the following components:
| Layer | Tools & Ideas |
|---|---|
| 🎥 Pose Estimation | MediaPipe, OpenCV, Skeleton Tracking |
| 🧠 Model Logic | TensorFlow or PyTorch for classification + metrics |
| 📊 Data Engineering | Custom pipelines to extract & label joint movement |
| 🎛️ App Layer | Streamlit for quick UI (MVP), then React/Flask for production |
| ☁️ Infrastructure | GitHub, Docker, AWS S3 for scaling later |
| 📈 Insights & Feedback | Angle calculations, comparison heatmaps, audio/text feedback |
In the early stages, I’ll work with public golf swing datasets and simulated motion data. But the long-term dream? Incorporating real-world video submissions from everyday users—closing the loop between model training and actual impact.
🔍 Why This Project Matters (and Why Now)
This isn’t just about golf—it’s about democratizing expert feedback using tech.
Imagine a high-school athlete in Nairobi, a retired golfer in Atlanta, or a beginner mom in Lagos… all receiving world-class coaching through AI. No private instructor. No expensive sensors. Just a smartphone and a personalized feedback loop.
For me, SwingSense is proof that:
- ML and data engineering can solve meaningful real-world problems
- Sports and science don’t have to be separate lanes
- And that a Black African woman can architect next-gen tools on a global scale.
🗺️ What You’ll Learn in This Series
If you’re following this series, expect a guided journey through:
- 🎯 Pose detection & human movement modeling
- 🔍 Golf swing feature extraction & biomechanics mapping
- 🤖 Training an ML model to assess swing quality
- 🛠️ Deploying a feedback engine via Streamlit
- 📦 Lessons on building, breaking, and refining as we go
It’ll be real, raw, and rigorously documented—with code snippets, visuals, personal insights, and aha moments along the way.
👋🏾 Let’s Connect (and collaborate)
Whether you:
- Love golf and want to explore the tech side,
- Are building your own AI/ML project and need a sister-in-code,
- Or you’re just curious how data science meets sports elegance—
…I’d love to hear your thoughts, questions, or even co-create something magical. 💬
🔮 Up Next: Pose Estimation in Action
In the next post, we’ll step into the swing—quite literally—by:
- Loading our first swing video
- Extracting joint landmarks using MediaPipe
- Visualizing motion frames in real time
- And preparing our data pipeline for training
So, grab your favorite club (or keyboard) and join me on this beautiful blend of logic, movement, and machine learning.
Until then, keep it precise. Keep it playful. And always… swing with purpose.