TMY3 GHI & DNI Weekly Aggregation

Read hourly solar weather data from TMY3 CSV files, compute weekly average GHI and DNI for each station, output to JSON.

Requirements

• Python 3.7+
• pandas
• matplotlib

Setup

# Install uv if needed
pip install uv

# Install dependencies
uv sync

Run

uv run solution.py

Expects tmy3.csv and TMY3_StationsMeta.csv in current directory.

Output:

• output.json (1,020 stations, 53 weeks each, ~5 MB)
• PNG visualizations for first 5 stations

Data Files

Download tmy3.csv and TMY3_StationsMeta.csv from Kaggle TMY3 dataset. Place in the same directory as solution.py.

Output Format

See output_sample.json for an example (3 of 1,020 stations). Full output contains all stations with the same structure:

[
  {
    "id": "<station ID>",
    "site_name": "<station name>",
    "coordinates": [<longitude>, <latitude>],
    "data": [
      {
        "timestamp": 1234567890000,
        "ghi": 120.5,
        "dni": 95.2
      },
      ...
    ]
  },
  ...
]