Dataset Viewer
Auto-converted to Parquet Duplicate
obs_id
int64
norad
int64
object
string
station
int64
start
string
intdes
string
frequency_hz
float64
station_lat
float64
station_lon
float64
station_alt_m
float64
time_mjd
list
freq_recv_hz
list
n_points
int64
13,494,755
64,879
Geoscan-6
4,394
2026-03-02
2025-155
436,938,900
33.893
130.849
10
[ 61101.25269371324, 61101.25269782846, 61101.25271223176, 61101.25271634698, 61101.25272457743, 61101.25273383669, 61101.252737951916, 61101.25274721118, 61101.25276058566, 61101.2527647009, 61101.25277190254, 61101.252772931344, 61101.252774988956, 61101.25278013299, 61101.2527811618, ...
[ 436957237.64865506, 436958416.7494797, 436957578.1459375, 436959156.9274675, 436961228.77555674, 436958355.6808197, 436958743.2984933, 436954143.82947, 436956087.74434483, 436956176.1497416, 436957935.4727688, 436957936.2810203, 436958619.7132139, 436955520.46315926, 436955399.5059312, ...
963
13,672,509
64,879
Geoscan-6
4,394
2026-03-27
2025-155
436,938,900
33.893
130.849
10
[ 61126.265601520674, 61126.26560563588, 61126.265606664696, 61126.2656076935, 61126.265608722315, 61126.26560975112, 61126.265610779934, 61126.26561180874, 61126.26561386634, 61126.265621068, 61126.26562209679, 61126.2656231256, 61126.26562415441, 61126.265625183216, 61126.26562621203, ...
[ 436938636.816565, 436933720.11767364, 436934245.79257894, 436938919.56754565, 436940505.3197384, 436934416.89085954, 436936217.79733247, 436936372.3483249, 436935031.4148704, 436940868.6817892, 436934520.13289, 436940298.6078683, 436938293.2920734, 436936553.1165541, 436941496.49185646, ...
1,499
12,229,173
64,879
Geoscan-6
40
2025-08-21
2025-155
436,938,900
52.834
6.379
10
[ 60908.15215560763, 60908.15216178046, 60908.15216589569, 60908.152170010915, 60908.15217309734, 60908.15219573109, 60908.15220704795, 60908.15221013438, 60908.152211163186, 60908.15221939364, 60908.1522296817, 60908.15223071051, 60908.15223996977, 60908.15224099858, 60908.152248200226, ...
[ 436942776.5258253, 436942780.9587462, 436944454.89252377, 436945806.83383054, 436948674.4691162, 436945013.93622524, 436947153.8787242, 436946252.24213946, 436944470.76866144, 436948666.7122873, 436942764.6849762, 436943605.5619407, 436943335.0840846, 436942589.888502, 436944448.8955999,...
1,525
13,482,168
64,879
Geoscan-6
2,940
2026-02-27
2025-155
436,938,900
47.02
15.263
410
[ 61098.12520307535, 61098.12522570911, 61098.12546130581, 61098.12546233462, 61098.12546336343, 61098.12546439224, 61098.125465421035, 61098.12546644985, 61098.125467478654, 61098.12546850745, 61098.125469536266, 61098.12547056507, 61098.12547159388, 61098.12547262269, 61098.12547365149, ...
[ 436948117.7545166, 436948115.83996683, 436948090.8276677, 436948120.84609467, 436948092.7705826, 436948092.90605754, 436948092.1005759, 436948091.0389641, 436948088.2912215, 436948090.15401036, 436948095.52860934, 436948086.3576995, 436948089.1236291, 436948091.21905595, 436948089.812550...
611
13,494,754
64,879
Geoscan-6
4,394
2026-03-01
2025-155
436,938,900
33.893
130.849
10
[61100.27006578067,61100.27007503993,61100.27007606874,61100.27007709755,61100.27008018395,61100.270(...TRUNCATED)
[436953554.29681593,436953593.03983593,436957421.07320505,436955617.2350507,436953195.88868266,43695(...TRUNCATED)
938
12,945,150
64,879
Geoscan-6
4,394
2025-12-13
2025-155
436,938,900
33.893
130.849
10
[61022.26942945858,61022.269433573805,61022.26943666022,61022.26943974664,61022.269442833065,61022.2(...TRUNCATED)
[436944794.2038963,436944558.0159915,436949401.0587482,436943993.6191145,436944543.5463252,436946244(...TRUNCATED)
929
13,474,892
64,879
Geoscan-6
2,940
2026-02-25
2025-155
436,938,900
47.02
15.263
410
[61096.53859064813,61096.53859167693,61096.53861945471,61096.53862048352,61096.53862151233,61096.538(...TRUNCATED)
[436947736.4512697,436947717.53126967,436947695.3374678,436947698.62527597,436947693.26887625,436947(...TRUNCATED)
995
12,229,044
64,880
Geoscan-1
3,856
2025-08-20
2025-155
435,973,500
54.523845
18.513308
56
[60907.547177935776,60907.54717999338,60907.547182051,60907.54718513742,60907.54718616622,60907.5471(...TRUNCATED)
[435976983.4404645,435976995.1715652,435977024.9796462,435977005.42893964,435975540.6873308,43597699(...TRUNCATED)
1,551
12,945,149
64,880
Geoscan-1
4,394
2025-12-13
2025-155
435,973,500
33.893
130.849
10
[61022.259083141354,61022.25908417016,61022.25909137181,61022.259092400614,61022.25909342943,61022.2(...TRUNCATED)
[436003558.9767553,436003551.8466217,436003547.82694846,436003545.4142328,436003553.0024164,43600353(...TRUNCATED)
966
13,763,252
64,880
Geoscan-1
4,394
2026-04-09
2025-155
435,973,500
33.895
130.84
10
[61139.2536425814,61139.25364566782,61139.253649783044,61139.25365184066,61139.25365698469,61139.253(...TRUNCATED)
[435986372.63795805,435988536.61826885,435986497.2850697,435983588.55715835,435984761.3926674,435981(...TRUNCATED)
1,491
End of preview. Expand in Data Studio

SatNOGS Doppler — near-identical cluster identification

Labeled Doppler tracks extracted from SatNOGS waterfall observations, for the task of identifying which member of a near-identical rideshare cluster produced a given pass. Built by satnogs-id, which wraps Cees Bassa's strf (rffit) — no trained model; the labels come from physics.

What's in it

One row per observation pass. The SatNOGS waterfall is Doppler-corrected, so each track is un-corrected back to the physical received frequency — freq_recv = f0 + offset − f0·range_rate/c — the curve rffit actually fits. This is non-circular: it does not depend on which TLE was applied.

column meaning
obs_id SatNOGS Network observation id
norad truth label — SatNOGS-assigned catalog object
object human name (e.g. Geoscan-4)
station SatNOGS ground-station id
start observation date
intdes launch international designator
frequency_hz nominal transmit frequency
station_lat, station_lon, station_alt_m receiver location
time_mjd per-point time (MJD) — variable length
freq_recv_hz per-point un-corrected received frequency (Hz) — variable length
n_points track length

Current contents

51 passes across two identified near-identical clusters (distinguish them by the intdes column), the second held out as a generalisation test. Identification is rffit's own identify_satellite_from_doppler.

  • Geoscan (2025-155) — 6 identical cubesats, 28 passes, 28/28 top-1 (all rank-1, CI 88–100%).
  • Tevel-2 (2025-052, held-out) — 9 identical cubesats, 23 passes, 20/23 top-1 (87%, CI 68–95%; all 3 misses are rank-2 near-ties). A different bus / band / geometry the method never saw — it generalises, but isn't a suspiciously-perfect 100%.

Usage

from datasets import load_dataset
ds = load_dataset("ryroeu/satnogs-id-doppler", split="train")
row = ds[0]
# row["time_mjd"], row["freq_recv_hz"] -> the Doppler track; row["norad"] -> truth label

To refit with strf, write each row as an rffit .dat (MJD freq_hz 1.0 site) plus a sites.txt line from station_lat/lon/alt_m.

Limitations

  • Two clusters so far; both amateur UHF cubesat rideshares. Wider orbit/band coverage is future work.
  • Truth is the SatNOGS-assigned identity, not independently decoded telemetry.
  • Tracks are auto-extracted (parabolic sub-bin peak + carrier window + MAD rejection); weak / low-elevation passes carry less Doppler curvature, so their margins are thinner.

Credit & license

Data from the SatNOGS network + DB (CC BY-SA 4.0). Identification engine: strf (Cees Bassa). Pipeline: satnogs-id.

Downloads last month
23