*The map that argues back*
| https://argos-geo.org
| github.com/argos-geo
: : [email protected]
Meet ARGOS-not a beach, not a great wine but an open source digital twin for the entire island- Built in public, with a full-time job, because someone has to.
1/4
48 hours of soak testing, passed.
Every road and building on the island now serves from production.
The stack is unchanged.
The twin never sleeps anymore.
: : argos-geo.org/map
48 hours of soak testing, passed.
Every road and building on the island now serves from production.
The stack is unchanged.
The twin never sleeps anymore.
: : argos-geo.org/map
We promised a pluvial term in v1.1.
Flood model v1.1 is live.
We promised a pluvial term in v1.1.
Flood model v1.1 is live.
The documented damage reads like a case file for the map we are building.
1/6
The documented damage reads like a case file for the map we are building.
1/6
It said 44 minutes.
I didn't want that answer, so I tried to break it.
Six bugs later, it still said 44 minutes.
It said 44 minutes.
I didn't want that answer, so I tried to break it.
Six bugs later, it still said 44 minutes.
Today the wildfire layer took its first independent exam.
We crossed 25 years of NASA FIRMS fire detections, 397 points across the island, with our risk classes.
This time, the model didn't get to mark its own homework.
Today the wildfire layer took its first independent exam.
We crossed 25 years of NASA FIRMS fire detections, 397 points across the island, with our risk classes.
This time, the model didn't get to mark its own homework.
Wildfire risk. Flood risk. Fire-response coverage.
The interesting part isn't only what the map shows, but how we know.
Every layer is bilingual and comes with its own evidence notes.
Wildfire risk. Flood risk. Fire-response coverage.
The interesting part isn't only what the map shows, but how we know.
Every layer is bilingual and comes with its own evidence notes.
In Peratata, a speedboat was deployed to rescue residents.
It was the worst flooding there in 60 years.
Afterward, I ran the ARGOS drainage model against the storm reports.
In Peratata, a speedboat was deployed to rescue residents.
It was the worst flooding there in 60 years.
Afterward, I ran the ARGOS drainage model against the storm reports.
For ARGOS, it’s more like the beginning.
The interesting part isn’t seeing a road, a building, a beach or a wildfire-risk layer on a screen.
It’s asking what happens when you connect them.
For ARGOS, it’s more like the beginning.
The interesting part isn’t seeing a road, a building, a beach or a wildfire-risk layer on a screen.
It’s asking what happens when you connect them.
But what happens when we add everything around it?
Settlements. Roads. Terrain. Land cover. Critical infrastructure. Evacuation points. Population. Accessibility.
Suddenly, we're not just looking at a risk layer.
But what happens when we add everything around it?
Settlements. Roads. Terrain. Land cover. Critical infrastructure. Evacuation points. Population. Accessibility.
Suddenly, we're not just looking at a risk layer.
It doesn't necessarily need to start as a huge platform.
It can start with an island, open data, a map, and a simple question:
What happens when local knowledge and open technology come together?
It doesn't necessarily need to start as a huge platform.
It can start with an island, open data, a map, and a simple question:
What happens when local knowledge and open technology come together?
ARGOS ran its wildfire risk index against 24 historical EFFIS fires (2017-2026) before this season started.
The model already knew where the risk concentrated. The burns validated it.
ARGOS ran its wildfire risk index against 24 historical EFFIS fires (2017-2026) before this season started.
The model already knew where the risk concentrated. The burns validated it.
ARGOS already uses Copernicus Sentinel-2 for vegetation stress and burn scar detection. Sentinel-3 adds thermal and ocean color. The stack gets stronger.
ARGOS already uses Copernicus Sentinel-2 for vegetation stress and burn scar detection. Sentinel-3 adds thermal and ocean color. The stack gets stronger.
NDVI on karst limestone is a liar. It gives the same value for stressed fir and bare rock.
NDVI on karst limestone is a liar. It gives the same value for stressed fir and bare rock.
GeoServer: The standard. Java. Heavy. XML configuration.
Martin: Light, fast, Rust. But vector tiles only. No raster queries. No custom API logic.
PostGIS: The database that happens to be a geospatial engine. SQL + Python. I speak both.
GeoServer: The standard. Java. Heavy. XML configuration.
Martin: Light, fast, Rust. But vector tiles only. No raster queries. No custom API logic.
PostGIS: The database that happens to be a geospatial engine. SQL + Python. I speak both.
ARGOS dark is built for the twin: navy land, gold contours, EMODnet bathymetry, Copernicus hillshade.
Satellite is Esri World Imagery. Use it for the real cross-check: is that building actually there? Does that road actually exist?
ARGOS dark is built for the twin: navy land, gold contours, EMODnet bathymetry, Copernicus hillshade.
Satellite is Esri World Imagery. Use it for the real cross-check: is that building actually there? Does that road actually exist?
climate.esa.int/en/Cross_ECV...
climate.esa.int/en/Cross_ECV...
I keep thinking about what that number means when the people who need help do not know the island, and the people who know the island do not know where the tourists are.
I keep thinking about what that number means when the people who need help do not know the island, and the people who know the island do not know where the tourists are.
616,000 hectares burnt in the EU this year.
Double the 20-year average.
Greece is in the Extreme fire danger zone for the next 7 days.
The map that argues back does not show you Greece. It shows you which ridge burns first.
616,000 hectares burnt in the EU this year.
Double the 20-year average.
Greece is in the Extreme fire danger zone for the next 7 days.
The map that argues back does not show you Greece. It shows you which ridge burns first.
12,217 roads.
22,013 buildings.
1,831 POIs.
197 beaches.
One PostgreSQL database holding an entire Greek island.
Click anything. A building, a beach, a road. It answers. Fire response time. Nearest port. As-the-ambulance-drives, not as-the-crow-flies.
#geospatial #data
12,217 roads.
22,013 buildings.
1,831 POIs.
197 beaches.
One PostgreSQL database holding an entire Greek island.
Click anything. A building, a beach, a road. It answers. Fire response time. Nearest port. As-the-ambulance-drives, not as-the-crow-flies.
#geospatial #data
Meanwhile I'm still trying to answer one question: if a fire starts on that ridge on Kefalonia, how long until the fire truck actually arrives?
Continental frameworks are important. Local granularity is *urgent*
Extreme heat could cost EU economies around €180 billion this year, roughly 1% of GDP and almost exactly the growth the bloc had been expecting for 2026, with France facing the heaviest blow of the big member states.
Meanwhile I'm still trying to answer one question: if a fire starts on that ridge on Kefalonia, how long until the fire truck actually arrives?
Continental frameworks are important. Local granularity is *urgent*
Watch our tutorial "Sea Surface Temperature Trends Under Climate Change" to learn how to analyse #CopernicusMarine data using Python 👉 https://bit.ly/4hZjgfF
📸: European Union, Copernicus Marine Service Data
Watch our tutorial "Sea Surface Temperature Trends Under Climate Change" to learn how to analyse #CopernicusMarine data using Python 👉 https://bit.ly/4hZjgfF
📸: European Union, Copernicus Marine Service Data
Hardest place to reach? Alighthouse on Vardiani islet that was built by the British in 1824, bombed by the Italians in 1942 and flattened by the 1953 earthquakes. Accessible only by boat.
#opensource #digitaltwin #kefalonia
Hardest place to reach? Alighthouse on Vardiani islet that was built by the British in 1824, bombed by the Italians in 1942 and flattened by the 1953 earthquakes. Accessible only by boat.
#opensource #digitaltwin #kefalonia
I wrote up the full pipeline, from downloading granules to the moment i realised NDVI 0.3 might mean just rocks.
The satellite sees what it sees. We'll figure out the context together. #geo
I wrote up the full pipeline, from downloading granules to the moment i realised NDVI 0.3 might mean just rocks.
The satellite sees what it sees. We'll figure out the context together. #geo
Full mobilisation and rightly so.
But how does the daily risk map reach municipalities? How are decisions actually made?
1/2
Full mobilisation and rightly so.
But how does the daily risk map reach municipalities? How are decisions actually made?
1/2