UTDQuake: University of Texas at Dallas Earthquake Dataset

A global earthquake dataset constructed from high-quality source and receiver metadata, including associated seismic phase picks across diverse station geometries.

Installation (utdquake)

pip install utdquake

Documentation

Full documentation for UTDQuake is available here:

You will see:

• QuickStart guide to get you up and running
• Detailed API reference
• Tutorials and example workflows

Dataset

Why this dataset matters?

Curated datasets of earthquake events and phase picks are essential for modern seismology, especially in the AI era. While waveform datasets have advanced earthquake detection, multistation picks provide complementary information crucial for phase association and earthquake location.

This dataset offers structured event catalogs, station metadata, and phase picks across networks, supporting reproducible research and the development of data-driven seismological methods.

What’s inside?

Directory	Format	Description
network/	*.parquet	Network metadata, one row per UTDQuake network (utdq_net).
events/	*.parquet	Earthquake events, one file per utdq_net.
stations/	*.parquet	Station metadata, one file per utdq_net.
picks/	*.parquet	Seismic phase picks, one file per utdq_net.
banks/	*.zip	ObsPlus EventBank datasets, one per utdq_net. Can be read directly using ObsPlus EventBank.

Naming: utdq_net, network and catalog

• utdq_net: the UTDQuake network, i.e. the monitoring network or project that reported the events and picks (e.g. tx, AEIC, R05). It is the key of the network table and of every file, and links the four tables. Catalogs built by research projects get an R code.
• catalog: the catalog or bulletin the events come from (e.g. ComCat, ISC Bulletin, ToC2ME). Network table only.
• network: the FDSN network code of the station that recorded a pick (e.g. AK), as in its seed_id. Stations and picks tables only.

Join picks to stations on (utdq_net, network, station) and picks to events on (utdq_net, event_id).

For details on the contents and schema of each dataset, please refer to the Hugging Face dataset viewer.

To get started, see the Quick Start section below, or click “Use this dataset” on the Hugging Face dataset page for example loading code.

Quick start

Basic Access

import utdquake as utdq

# dataset overview 
dataset = utdq.Dataset(das=False) #das=True for DAS data access.
print(dataset)

# network level
network_data = dataset.networks
print(network_data)

dataset.plot_overview(savepath="utdquake.png")

Network Data

# load network 
network = dataset.get_network(name="tx")   # by utdq_net
print(network)

# events
events = network.events
print(events)

# stations
stations = network.stations
print(stations)

# picks
picks = network.picks
print(picks)

Event Bank

Check ObsPlus EventBank for more details.

# get event bank
ebank = network.bank # 

# Example: Filter by event_id
ev_ids = events["event_id"].iloc[:5].tolist()
cat = ebank.get_events(event_id=ev_ids)
print(cat)

# Example 2: Other filter (check obsplus.EventBank for more details)
cat2 = ebank.get_events(minmagnitude=4.3)
print(cat2)

Plot

# get Obspy Event
network = dataset.get_network(name="tx")
network.plot_overview(savepath="overview.png")
network.plot_uncertainty_boxplots(savepath="uncertainty_boxplots.png")
network.plot_station_location_uncertainty(savepath="station_location_uncertainty.png")
network.plot_stats(savepath="stats.png")
network.plot_pick_histograms(savepath="histograms.png")
network.plot_pick_stats(savepath="pick_stats.png")

Thanks

Thanks to the UT Dallas HPC team for providing the computational resources for this dataset.

We also thank the seismology and AI communities for their work in earthquake research, and Hugging Face for hosting and sharing open datasets.

We welcome feedback and contributions!

Authors

Emmanuel Castillo University of Texas at Dallas
Nadine Ushakov University of Texas at Dallas
Marine Denolle University of Washington

---

license: gpl-3.0