Machine Learning Coursework
Weekly Price Forecasting
A time-series coursework exercise using weekly avocado prices across regions and product types. I constructed next-week targets and lag features, used a chronological train/test split, and compared Random Forest predictions with a baseline that reuses the current price.
Period
Jan - Apr 2026
Tools / Tech
PythonPandasScikit-learnRandom ForestLag Features
Why I built it
I wanted to understand how forecasting differs from a random train/test split, especially when the dataset contains several related time series.
Links
What it includes
- Groups observations by region and product type before creating lagged prices.
- Adds calendar features and compares train/test R-squared with a persistence baseline.
What I worked on
- Wrote the data exploration, grouped lag construction, and Random Forest experiment myself within the course notebook.
- Checked timestamp spacing and compared the model with a simple forecast before interpreting its score.
What I Learned
- Practiced aligning features and future targets within each series.
- Learned how a strong simple baseline changes the way a forecasting result should be interpreted.