
In 2025, more than eight thousand earthquakes of magnitude 4.5 and above shook the Earth. Put them on one map and the map draws a ring by itself: the Pacific Ring of Fire. In the next hour you will make this map yourself. This lesson shows the plan and the three traps we will meet on the way.
Resources: c0-data.zip (all course data: Natural Earth layers, a copy of the USGS 2025 earthquakes and a ready GeoPackage). The 00_README.txt inside explains which lesson uses which file.
The QGIS download page offers two versions: the latest release and the long-term release (LTR). We pick one on purpose, then install it on Windows, macOS or Linux.
Lesson note: the course was recorded with QGIS 4.2. In a newer 4.x release the steps are almost always the same; if a button moves, a note is added here. Download: qgis.org/download
The window looks crowded, but five parts are enough to start: the map canvas, Browser, Layers, the toolbars and the status bar. We bring back a closed panel and find anything with the Locator.
Natural Earth is public domain. We download countries and the ocean at medium scale, add them to QGIS without unzipping, and save the project. We also see why a shapefile is not one file and why borders follow de facto control.
Lesson note: naturalearthdata.com → Downloads → Medium scale (1:50m) → Cultural: Admin 0 – Countries, Physical: Ocean. The same files are in c0-data.zip (Lesson 1 resources).
Latitude, longitude and EPSG codes. We switch the project to Equal Earth and put the Pacific in the middle. Trap one: stripes across the whole screen at the seam, why they appear and how to fix them.
Codes used: EPSG:4326 (WGS 84 latitude/longitude), EPSG:8857 (Equal Earth), EPSG:8859 (Equal Earth Asia-Pacific). Ready layers cut at the seam are in ring_of_fire.gpkg inside c0-data.zip.
Behind every shape is a row. We open the attribute table, sort, select and identify a single country. Trap two: the population in the table is from 2019. Checking the year of a number is the habit this lesson builds.
We add the USGS live feed of the past week's earthquakes straight from a web address. Trap three: quarry blasts and ice quakes are in the list too. We filter them out and see why the time column has 13 digits.
Live feed address (Add Vector Layer → Protocol → HTTP): https://earthquake.usgs.gov/earthquakes/feed/v1.0/summary/all_week.geojson
Counts in the video are from the recording day (8 Nov 2026); yours will differ.
The yearly earthquake file is really a question written as a web address. We read it piece by piece and change magnitude, dates and area to ask USGS our own question. The 20,000-event limit per query, and why you should keep the address.
The course's 2025 query: https://earthquake.usgs.gov/fdsnws/event/1/query?format=geojson&starttime=2025-01-01&endtime=2026-01-01&minmagnitude=4.5
A copy downloaded on 8 Nov 2026 is in c0-data.zip (usgs_2025_m45.geojson).
A magnitude 4.5 and an 8.8 looked exactly the same. Each whole step is about 32 times more energy. With graduated size, opacity, drawing order and legend precision we make the map honest.
Depth is not in the attribute table: it is the third number of the geometry. We find it with Identify, read it with the z($geometry) expression and style a copy of the layer in three depth classes. About 81% of earthquakes are shallow.
Eight thousand labels is noise, not a map. With rule-based labeling we write only the sixteen earthquakes of magnitude 7 and above, with a halo. The largest of 2025: Kamchatka, 29 July, M8.8.
A screenshot is not a map. In a print layout we add the map, a title, a legend and a source line, and see why a scale bar and a north arrow mislead on a world map.
Source line example: Data: USGS (downloaded 8 Nov 2026), Natural Earth. Projection: Equal Earth.
Grid, frame and type hierarchy. A curved grid is not a mistake: it is the projection made visible. Small changes make the page readable.
We export the map as PDF and as a 300 dpi PNG (about 3,500 pixels wide on A4). Then we package the whole project with its layers into one GeoPackage file that anyone can open.
The three traps on one card, and back to the first lesson. A final task: build the USGS address for 2024 yourself and export a second map the same way. Is the ring the same?
The task's address: https://earthquake.usgs.gov/fdsnws/event/1/query?format=geojson&starttime=2024-01-01&endtime=2025-01-01&minmagnitude=4.5
This course contains the use of artificial intelligence.
In 2025, the Earth shook more than eight thousand timesPut every earthquake of magnitude 4.5 and above on one map and the map draws a ring by itself: the Pacific Ring of Fire. In this short course you make that map yourself, starting from an empty screen. At the end you have a one-page PDF and PNG map with a title, legend, grid and source line, plus a QGIS project you can share as a single file.
The software is free and open source: QGIS 4. The data is free and open: Natural Earth and the US Geological Survey (USGS). No account, no credit card and no programming needed.
What you will doA map can look convincing and still be wrong. On the way we meet three traps on purpose: a projection that tears the map apart, a population figure from 2019, and dots that are not earthquakes at all (quarry blasts, ice quakes). You learn how to spot each one and how to fix it.
How it worksEach lesson teaches one skill and adds something visible to the map. Menu paths are shown on screen. Some lessons have a "pause and try" task. All course data is in the first lesson's resources as one zip file; if you get stuck, continue from the ready GeoPackage inside.
VersionThe videos were recorded with QGIS 4.2 in English. In a newer 4.x release the steps are almost always the same.
Use of AIThe narration of this course uses an AI-generated voice. AI tools were also used to help prepare the script and the visuals. All demos were recorded in real software with real data, and the instructor reviewed all content.