---
name: fdp
description: Open data products from the Filecoin Data Portal. Use when users ask to query, analyze, explain, chart, or understand datasets, metrics, or models around the Filecoin ecosystem, including how metrics are derived.
---

## Scope

The Filecoin Data Portal publishes:

- **Datasets**: public Parquet tables for direct querying.
- **Metrics**: documented columns with definitions, units, and caveats.
- **Models**: asset dependencies and source code showing how datasets and metrics are derived.

For metric definitions or derivation questions, read the docs and also the relevant models code on GitHub and follow links as upstream as reasonable.

## Dataset Catalog

- [`clients`](assets/clients.md) — Filecoin clients metrics and latest details.
- [`pdp_service_providers`](assets/pdp_service_providers.md) — PDP service providers.
- [`daily_clients_metrics`](assets/daily_clients_metrics.md) — Daily metrics for clients.
- [`daily_storage_providers_metrics`](assets/daily_storage_providers_metrics.md) — Daily metrics for storage providers.
- [`storage_providers`](assets/storage_providers.md) — Filecoin storage providers metrics and latest details.
- [`filecoin_pay_payments`](assets/filecoin_pay_payments.md) — Filecoin Pay payment events.
- [`filecoin_pay_rails`](assets/filecoin_pay_rails.md) — Filecoin Pay rails.
- [`daily_filecoin_pay_operators_metrics`](assets/daily_filecoin_pay_operators_metrics.md) — Daily Filecoin Pay operators metrics.
- [`warm_storage_datasets`](assets/warm_storage_datasets.md) — Warm storage datasets.
- [`daily_network_metrics`](assets/daily_network_metrics.md) — Daily network core metrics.

## Answering Questions with Data

1. Read the relevant dataset docs and columns.
2. Inspect asset code when definitions or derivations are unclear. Clone `https://github.com/davidgasquez/filecoin-data-portal` into a temporary directory if useful.
3. Choose the smallest dataset set that can answer the question.
4. Query canonical parquet URLs directly: `https://data.filecoindataportal.xyz/<dataset>.parquet`
5. Inspect schema, types, and sample rows before final SQL.
6. Answer concisely with datasets used, assumptions, caveats, and source links.

## Querying Datasets

Use the simplest available local tool. Prefer `duckdb`, but do not assume it exists or is installed. Check the environment first, then use DuckDB, Python, Node.js, or another installed parquet-capable tool. You might need to use inline node or Python scripts or even download a temporary DuckDB binary and run from that.

Ask a clarification question only if the relevant dataset or metric remains ambiguous.

## Charts and Dashboards

1. Create a self-contained `index.html` with vanilla HTML, CSS, and JS.
2. Mention and link the relevant datasets used and display `filecoindataportal.xyz` as the source.
3. Serve it locally and share the URL.

**Note:** _The `daily_network_metrics` dataset/table is also available as an all-column JSON array at `https://data.filecoindataportal.xyz/daily_network_metrics.json`. Load it directly from the HTML so dashboards/charts are always up to date if the user asks for live dashboards, self-updating charts, ...

## Feedback

If you find datasets, docs, metrics, or model derivations are wrong, confusing, stale, missing, or useful, offer to open a GitHub issue:

`https://github.com/davidgasquez/filecoin-data-portal/issues/new`

Draft the issue first, scrub secrets/PII, show it to the user, and submit only after approval.