About SeismoSight

Motivation

SeismoSight started as my personal automation project. I've always been curious about what's actually going on with earthquakes around the planet, but between family, work, and trying to have some kind of life, I just don't have the time (or patience) to manually dig through raw data every day.

So I built this site to do the boring heavy lifting for me — automatically collecting the data, processing it, and turning it into something simple and readable. What began as a personal tool to satisfy my own nerdy curiosity eventually turned into something I decided to share publicly.

Who this is for

SeismoSight is for anyone who's curious about earthquakes and wants a simple, clear picture of what's happening — without needing to be a seismologist or spend hours digging through raw data.

Whether you live in a seismic area, follow global events, love earth sciences, or are just a naturally curious person, this site is built for you.

New locations & features

I'm keeping this page short, so I'm putting updates about new locations and features into the blog instead. There's already one kick-off article there, and more will appear as I add new stuff over time.

You can find them in the blog section.

Data sources

All the earthquake data comes from two big public sources — USGS and EMSC. I pull it through the FDSN network, which is basically the standard (and most reliable) way to get this kind of data.

Nothing is scraped or invented — it's all real, openly available data from the official providers.

AI content

Because I don't have a seismology background and limited time, the site automatically calls AI to generate geological and historical context for new swarm pages and major earthquakes. It works the same way I personally use AI when I want to quickly understand something — just at scale and in real time.

I try to keep the prompts grounded in real data. You'll see disclaimers on AI-generated content because regulations require them. The actual earthquake numbers always come from USGS and EMSC.

Current limitations

Here are a few honest limitations of the site:

Data is refreshed every hour, so there can be up to a one-hour delay between an earthquake and when it appears on the site.
EMSC's data service occasionally goes down. When this happens, I pause updates completely rather than risk showing incomplete or misleading data. These outages are usually short, but they can sometimes last longer. I'm looking for a backup solution.
Historical data only goes back to January 1, 2000. This was a deliberate choice because earlier records (especially from EMSC) are much less complete. As a result, some statistical views may give a slightly misleading impression of increasing seismic activity over time.

Channels to follow

New events are automatically posted to Telegram, Reddit, and X. All links are available on the Follow page.

Wrapping up

SeismoSight started as my personal project to satisfy my own curiosity about earthquakes while dealing with real life — family, work, and limited time. Over time it grew into something I'm happy to share with others.

Thanks for stopping by. If you have feedback, ideas for new locations, spot something weird, or just want to chat about seismicity — feel free to reach out on X or Reddit, or .

Happy to keep exploring this restless planet together.

About the Ongoing News / Monthly Pages

News and summary pages on SeismoSight are organized by calendar month. Each monthly page consists of three main sections:

Ongoing Events

This section provides timely highlights of significant seismic activity during the month, including:

Short notes on ongoing M6.0+ earthquakes.
Swarm start and end events
Daily summaries: Concise, AI-generated overviews grounded in the actual data. Each summary covers major earthquakes, swarm developments (starts/ends), and notable divergences in daily seismicity relative to historical norms. Accompanying each summary is a table of daily earthquake statistics, including comparisons to historical mean averages.

To account for the gradual increase in the number and sensitivity of seismic detectors over time, daily mean averages are calculated using a rolling 5-year window based on recent data. This ensures fair and realistic comparisons.

List of M5.0+ Earthquakes

This section presents all earthquakes of magnitude 5.0 and above, directly sourced from USGS and EMSC catalogs. Events are grouped by day for easy reference.

List of Swarms

This section includes all earthquake swarms that intersect or overlap with the corresponding month (based on the swarm detection methodology).

Upcoming Improvements

Weekly summary sections.
At the beginning of each new month, an aggregated summary of the previous month.

These pages combine raw seismic data with human-readable context and statistics to help you quickly understand both significant events and broader seismicity patterns. All AI-generated content includes appropriate disclaimers, and primary data always comes from authoritative sources (USGS/EMSC).

Swarms

According to Wikipedia, an earthquake swarm is "a sequence of seismic events occurring in a local area within a relatively short period."

Since I originally built the SeismoSight engine for my own personal use without worrying about rigid scientific templates, I ended up developing a completely proprietary approach to swarm classification and criteria.

Swarm detection methodology

The current methodology uses an algorithm that searches for any possible constellation of earthquakes meeting defined swarm criteria (e.g., minimum event count, rate, and magnitude characteristics within a time window).

To prioritize historically targeted dense and localized swarms, the algorithm constrains each initial candidate swarm to events falling inside a fixed-radius circle. When multiple candidate circles overlap or intersect, the algorithm selects the single most dense and powerful candidate based on a composite score (incorporating event count and density).

Once an event is included in a swarm, it remains there permanently. As a result, the actual spatial area of an established swarm can grow over time and exceed the initial fixed-radius circle.

Known limitation / side effect: For widespread or elongated swarms, some qualifying events may still fall outside the selected newborn swarm's circle and thus be excluded. In some cases (particularly observed in the Hawaii region), a single coherent large swarm can be fragmented into multiple nearby smaller swarms, even though the events are physically related as part of one larger process.

Format of swarm's id

Format: [type][date].[index]

Types:

S - Swarm When at least 24 earthquakes of any magnitude are registered within any 15 km radius circle in a rolling 24-hour window, this type of swarm is triggered. The swarm is considered closed when the number of quakes in the same 24-hour rolling window drops to 2 or fewer. This one was the first type I created, and it was originally aimed at detecting lava-related swarms. In reality, it also catches aftershock sequences and industrial quakes really well.
PS - Powerful Swarm When at least 5 earthquakes of M5.0 or higher are registered within any 100 km radius circle in a rolling 24-hour window, this type of swarm is triggered. The swarm is considered closed when the number of quakes in the same 24-hour rolling window drops to one or fewer. This type was originally designed to catch potential foreshocks of strong earthquakes.
VS - Volcanic Swarm Same as the first type (S), but the center of the circle must also have a volcano within 5 km or less. This variant was supposed to detect eruption-related swarms.
SVS - Super-Volcanic Swarm Same as type S, but it also requires at least one quake inside a supervolcano area. I created this one specifically to monitor activity in big supervolcano calderas.
SVPS - Super-Volcanic Powerful Swarm Same as type PS, except it also needs at least one earthquake inside a supervolcano area. Interestingly, I haven't seen this type trigger even once since 2000.

Date: YYYYMMDD of the first earthquake that started the swarm.

Index: serial number of this swarm type within the day (starts from 1).

Examples:

S20260402.3 - Regular swarm. First quake hit on April 2, 2026 (UTC). This was the third swarm of type S that began on the same day.

SVS20240101.1 - Supervolcanic swarm. The first earthquake occurred on January 1, 2024 (UTC). It was the first swarm of this type on that day.