AruneshDev/Automated-Trading-System-Kalshi-Weather-Model — reverse-engineered prompt

Build me a simple end to end weather trading project that predicts the next day’s maximum temperature for New York City, Miami, Austin, and Chicago using the past few years of weather data, then uses those predictions to automatically place demo trades on Kalshi weather markets. I want it to feel like a student machine learning project that actually works, not a huge production system.

Please include the data loading and cleanup, training a model to predict daily max temperature, basic tuning to improve error, and clear charts showing predicted versus actual temperatures with mean squared error for each city. Then use the model’s daily predictions to choose the matching Kalshi temperature range market and place trades through the API.

Keep the workflow easy to run from notebooks, with one part for training and one part for automated trading. Save predictions and trade results so I can review settled outcomes later. If anything about the trading API has changed, look up the current docs online and wire it up in the simplest working way.