Services
Discovery, Desktop and
Mobile, UX/UI Design, Deploy
Date
Feb, 2022
Techologies
Node.Js
Flutter
Figma
What is Forest Inventory?
Here’s the Forest Inventory link
Arbolado urbano refers to the network of trees that shape the public spaces of a city — including streets, sidewalks, parks, and plazas. Urban trees are a vital part of the city’s ecosystem: they provide shade, improve air quality, regulate temperature, and contribute to the well-being and quality of life of citizens.
About the Municipality of San Isidro
The Challenge
San Isidro is an Argentinean city that faced a critical challenge in managing its urban trees: there was no centralized and reliable information about their location, species, health condition, or maintenance. This lack of structured data made planning inefficient, slowed down emergency response, and limited informed decision-making in environmental policies. A recent storm highlighted the urgency of having an accurate, centralized system to manage the city’s trees — a key step toward building safer, greener environments for the Citizens of the Future.
Our Solution
Rocking partnered with Ciudadanos del Futuro to launch a digital census of the urban forest using DeepForest, a machine learning algorithm specialized in tree crown detection.
- Iteration 1 → Initial implementation with DeepForest reached ~60% accuracy.
- Iteration 2 → Using Label Studio, more than 3,000 trees were manually labeled to improve training data. Accuracy increased to 79%.
- Iteration 3 → Introduced negative samples (non-tree elements) and fine-tuned thresholds and parameters. Accuracy further improved, moving closer to the 80–90% target.
Additional features:
- Crown size estimation delivered in GIS-ready format (Excel/CSV).
- Integration into Rocking Framework, providing each municipality with its own tenant, user roles (admin, editor, viewer), and map-based visualization.
- Editors can validate detections, mark errors, and add new trees.
- Public-facing web portal for citizens, branded with the municipality’s identity.
Impact
Achieved significant improvement in detection accuracy through iterative training, reaching close to the 80–90% target.
- Generated the first centralized digital inventory of San Isidro’s urban forest.
- Allowed better planning, monitoring, and early intervention in case of emergencies or tree health risks.
- Delivered a scalable multi-tenant platform for municipalities to manage their trees.
- Increased citizen engagement through a public map where residents can explore the city’s trees.
