# Day-Ahead Wind Speed Forecasting for Nord Pool Bidding

Pluto notebook accompanying the individual weather data paper for *Introduction to Business Analytics* (EBC1042), Maastricht University School of Business and Economics.

The paper proposes a day-ahead wind speed forecasting tool for the Alajärvi Wind Farm Complex (Ilmatar Energy), using daily FMI station data from Meteostat.

## Models

| Model | Description |
|-------|-------------|
| M0 | Persistence baseline (tomorrow = today) |
| M1 | Static OLS on Möksy plus the four inner-ring stations (10 lags) |
| M2 | Rolling 365-day OLS on Möksy plus inner and outer rings (18 lags) |

All models predict day *t* from observations at *t − 1* and *t − 2*. Calibration runs January 2021 to October 2025, and the holdout runs October 2025 to October 2026.

## Running the notebook

Requires [Julia](https://julialang.org/downloads/) and [Pluto.jl](https://plutojl.org/).

```bash
julia -e 'using Pluto; Pluto.run(notebook="Paper.jl")'
```

Pluto installs the required packages (GLM, DataFrames, LinearAlgebra) automatically on first run.

## Data

Daily wind speed observations from FMI synoptic stations, retrieved via [Meteostat](https://meteostat.net/). Place the data files in the project folder as described in the notebook.

## Files

- `Paper.jl`: the Pluto notebook with the full analysis
- `README.md`: this file
- `LICENSE`: license terms

## Author

Mikka Anton Petterson