Overview
The Rambus Project focuses on improving forecast accuracy for semiconductor market metrics such as ASP, revenue, demand, and supply. Using Python, pandas, and matplotlib, I developed a time-series model that reduced prediction errors by up to 20%. The analysis incorporated Omdia market data and explored multi-quarter forecasting trends to support Rambus’ strategic planning.
Error Themes Across Forecast Horizons
This chart visualizes how prediction errors for ASP, revenue, demand, and supply change as the forecast window shortens.
Data Preprocessing and Model Training
This code snippet shows the cleaning, feature engineering, and preparation steps used before model fitting.
Revenue Forecast Performance
A closer look at revenue-specific predictions, highlighting how the model accuracy improved for short-term forecasts.
