Observable Notebooks Data Loaders
4 days ago
- #Observable Notebooks
- #Data Loaders
- #Node.js Python
- Data loaders in Observable Notebooks run at build time for static data preparation, ensuring consistency and performance.
- Supported languages for data loaders include Node.js and Python, with potential for more in the future.
- Data loaders support various text-based (e.g., JSON, CSV) and binary formats (e.g., Arrow, Parquet), as well as image and HTML formats.
- Outputs from data loaders are saved in a local `.observable/cache` directory, with stable snapshots that update only upon re-running the cell.
- Node.js data loaders require Node.js 22.12+ and have security restrictions limiting file access to the notebook's directory.
- Python data loaders require Python 3.12+ and can utilize virtual environments, but packages must be manually installed.
- Examples include a Python cell printing a greeting and a Node.js cell fetching npm download statistics for Observable Plot.