---
license: cc-by-4.0
pretty_name: Constellation Demo Fleet
language:
  - en
tags:
  - space
  - satellites
  - ground-stations
  - telemetry
  - link-budget
  - snr
  - forecasting
size_categories:
  - n<1K
configs:
  - config_name: links
    data_files: links.csv
    default: true
  - config_name: ground_stations
    data_files: ground_stations.csv
  - config_name: satellites
    data_files: satellites.csv
  - config_name: demand_pools
    data_files: demand_pools.csv
---

# Constellation Demo Fleet

A ready-made satellite and ground-station fleet from [Constellation Space](https://constellation.space): 6 ground stations, 39 satellites, 21 ground-antenna links with link-budget state, and 4 traffic demand pools. It is the same fleet the Constellation API loads with `load_demo_fleet`, so you can explore it here and then run forecasts on it in the [Constellation Platform](https://platform.constellation.space).

This release is a **snapshot**: each entity's newest values as of `2026-10-10T23:07:10Z`. A time-series release (two hours of 15-second SNR telemetry and seven days of ten-minute demand) will follow.

## Files

| Config | File | Rows | What it holds |
|---|---|---|---|
| `links` (default) | `links.csv` | 21 | One ground antenna to one satellite: pointing, geometry, RF configuration, and observed SNR |
| `ground_stations` | `ground_stations.csv` | 6 | Station name, WGS84 latitude and longitude, and the demand pool it serves |
| `satellites` | `satellites.csv` | 39 | NORAD catalog id and name (O3b, O3b mPOWER, Globalstar, PROBA-2, PROBA-3, Sentinel-1C, BIOMASS) |
| `demand_pools` | `demand_pools.csv` | 4 | Latest ten-minute offered traffic per pool |
| | `topology.json` | 70 entities | The same snapshot in the shape the API's topology endpoint returns |

## Columns

### `links.csv`

| Column | Unit | Meaning |
|---|---|---|
| `entity_id` | | `<station>__ant<n>`: one antenna at one station |
| `source` | | The ground station (`ground_stations.entity_id`) |
| `target` | | The satellite (`satellites.entity_id`) |
| `observed_at` | UTC | Time of the newest sample |
| `azimuth_deg` | degrees | Pointing azimuth, clockwise from true north, in the station's local frame |
| `elevation_deg` | degrees | Pointing elevation above the local horizon |
| `slant_range_km` | km | Line-of-sight distance from antenna to satellite |
| `frequency_ghz` | GHz | Carrier frequency |
| `bandwidth_mhz` | MHz | Channel bandwidth |
| `eirp_dbw` | dBW | Effective isotropic radiated power of the transmitter |
| `g_t_db_k` | dB/K | Receive figure of merit (antenna gain over system noise temperature) |
| `snr_db` | dB | Observed signal-to-noise ratio |

### `ground_stations.csv`

`entity_id`, `name`, `lat_deg` and `lon_deg` (WGS84 degrees), and `pop_id`, the demand pool the station serves.

### `demand_pools.csv`

`pool_id`, `interval_start` (UTC, start of a ten-minute interval), `inlet_bytes` (offered bytes in that interval), and `evidence_kind`.

## Things to know

- **Demand is synthetic.** Every pool row has `evidence_kind = synthetic_offered`: it is generated traffic for exercising demand forecasts, not measured customer traffic.
- **Not every link closes.** `SITE-9c5100cd2d6c__ant0` to PROBA-3 sits near 48,800 km slant range at -12.5 dB SNR, which is realistic for a spacecraft near the far point of a highly elliptical orbit and is kept as an example of a link that does not close.
- **Links come and go.** Antennas that are not tracking a satellite stop reporting, so `observed_at` varies by link.
- **Satellite names and NORAD ids are public catalog identifiers.**

## Use it

```python
from datasets import load_dataset

links = load_dataset("constellationspace/demo-fleet", "links", split="train")
print(links.to_pandas().sort_values("snr_db").head())
```

Run forecasts on the same fleet with the API (create a key in the [Platform](https://platform.constellation.space), then load the fleet from the Dashboard or an MCP client):

```bash
curl -H "x-api-key: $CONSTELLATION_API_KEY" https://api.constellation.space/topology
```

## License

[CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Credit "Constellation Space" when you use it.