📈 Predicting Stock Price Movements with ML & Technical Analysis

Stock traders need reliable, data-driven signals to make profitable decisions. Traditional chart analysis is subjective and inconsistent — what if we could automate it with machine learning?

🔍 The Problem

Traders struggle with:

• Predicting price direction (up or down movement)

• Forecasting precise price targets for position sizing

• Reducing human bias in technical analysis

• Processing hundreds of stocks simultaneously

• Backtesting strategies with consistent signals

💡 The Solution

I built an end-to-end machine learning system that:

• Uses XGBoost for price direction classification (up/down prediction)

• Predicts exact closing prices using regression models

• Engineers 11+ technical indicators from OHLCV data

• Serves predictions through a Flask REST API

• Is fully containerized with Docker

• Deploys on Kubernetes for scalable inference

• Includes 1,219 labeled candlestick charts for training

🚀 Tech Stack

• ML: XGBoost (classification + regression)

• Backend: Flask + Gunicorn

• DevOps: Docker + Kubernetes + minikube

• Data: Synthetically generated OHLCV with human annotations

Full project here:

🔗 github.com/AlexHuc/OHLC…

This project was created as the capstone 1 assignment for the Machine Learning Zoomcamp by Alexey Grigorev, and it helped me strengthen my skills in ML modeling, deployment, containerization, and MLOps.

#mlzoomcamp #machinelearning #learninginpublic #datascience