Ellen Gomes

GIS and environmental data
Python · remote sensing · forest data

M.Sc. Forest Information Technology
B.Sc. International Forest Ecosystem Management

See my work Portrait of Ellen Gomes

Work

I turn raw environmental data into maps and answers: from drone flights and LiDAR point clouds to Python workflows on open government data. My focus is forests and ecosystems, and I work with Python, QGIS, ArcGIS, R and Google Earth Engine. The newest projects are at the top. Open a map to explore it, or click an image to see it in full size.

Automated Forest Indicators for Graubünden

How much of a municipality is covered by trees, and how tall is the forest? A Python workflow calculates canopy cover, mean forest height and upper canopy height for five Graubünden municipalities from the national forest inventory's canopy height model, and shows the results on an interactive map.

Data: Vegetationshöhenmodell LFI 2025 (WSL/BAFU), swissBOUNDARIES3D (swisstopo) · Tools: Python, rasterio, GeoPandas, NumPy, Folium · Open full map → · Code on GitHub →

Canopy height in Chur at 1 m resolution

A Closer Look: Canopy Height in Chur

The 1 m canopy height model behind the analysis. Forests on the valley slopes reach 20 to 35 m, while the Rhine valley floor and the alpine area above the tree line stay near 0 m. This is why canopy cover is only 42 % in Chur.

Data: Vegetationshöhenmodell LFI 2025 (WSL/BAFU) · Tools: Python, rasterio, Matplotlib

Can You Trust the Location? Checking umwelt.info Search Results

When you search the federal environmental data portal umwelt.info for forest data in Brandenburg, do the results really lie in Brandenburg? A Python workflow downloads 200 search results from the umwelt.info API, draws the bounding box of each dataset and compares it with the state boundary. About two thirds lie in Brandenburg, one in four lies near it (mostly Berlin and map sheets on the border), and 13 lie somewhere else.

Data: umwelt.info API, VG250 and TopPlusOpen (BKG) · Tools: Python, GeoPandas, Shapely, Folium

Open full map → · Code on GitHub →

LiDAR sun exposure simulation on roofs in Warsaw

Which Roofs Are Best for Solar Panels?

Using a LiDAR point cloud, sun exposure was simulated over 24 hours. The roofs that receive the most hours of light are the best candidates for installing solar panels.

Data: Geoportal Poland, Warsaw (private project) · Tools: ArcGIS, CloudCompare

Map of CO2 emissions from forest fires in northern Brazil

CO₂ Emissions from Forest Fires in Brazil

How much CO₂ did forest fires release in Brazil in 2018, and where? This map shows the emissions in the northern region, revealing where fires had the biggest climate impact.

Data: Global Forest Watch (GFW) · Tools: Python, NumPy, Pandas, Jupyter Notebook, Folium

Drone photogrammetry map of a forest regeneration area in Poland

Counting Young Trees from the Air

Photogrammetry from DJI Phantom 4 drone images of a forest regeneration area. The result supports counting the newly planted nursery trees.

Private project, Poland · Tools: Pix4D, ArcGIS

LiDAR point cloud of city trees near buildings, visualized in R

Where Are the Trees? LiDAR for Urban Tree Safety

A LiDAR point cloud visualized in R to measure the distance between trees and nearby buildings in the city, supporting safety assessments of tree condition.

Data: Geoportal Poland · Tools: CloudCompare, R (lidR package)