Blog 4 | Cognitive Deforestation Prevention | Deloitte Impact Foundation


Blog 4: Enabling action

Blog 4/4 from the series about the Deloitte Impact Foundation initiative ‘Cognitive Deforestation Prevention’ that aims to prevent illegal deforestation by building an artificial intelligence solution that predicts where illegal deforestation will happen

Bas Wagenmaker is part of the Visual Analytics team within Deloitte Consulting and has seen a wide variety of clients and analytics projects throughout his career. Working on the WWF’s Early Warning System as part of this Deloitte Impact Foundation initiative, has been one of the most rewarding projects he has done so far. In his blog he shares how he has developed a custom web application to enable local users to take action.

From data to insights

Bas: “An accurate and well-performing model is key in predicting illegal deforestation. However, without the right means to deliver these predictions to the end-user, the insights might lose their value and effective prevention may be significantly obstructed. Therefore, to navigate through the predictions and support follow-up on those predictions, an interactive and intuitive user interface with visualizations of the predictions and other relevant data was needed. 

User-interface Early Warning System

Scalable and customizable technology

When we joined the team, a proof of concept of the idea was already in place. One of our main challenges was to take it from a prototype to a fully customizable and scalable solution. In order to make this possible, we decided to create a custom user interface by leveraging state-of-the-art technologies and the power of the cloud. As a backbone for the development of this web application, we used open-source software like React and Node.js. 

To create an interactive map that allowed for multiple data layers created from huge data sets, we used Mapbox GL JS, a library that effectively improves performance by using technology like vector tiles. Vector tiles are especially beneficial in situations where internet connectivity is suboptimal and the amount of data that is downloaded needs to be limited. The generated vector tiles are relatively small of size, but more importantly they are only downloaded and rendered when they are needed for the specific area and detail level of the map that is in view. Mapbox also makes use of WebGL, which is a JavaScript API that allows for GPU-accelerated usage of image processing, thus not only using the CPU of a user, but also their video card to improve performance. The web application is deployed on Amazon Web Services (AWS) to further ensure scalability.

Various features were implemented to allow the user to effectively navigate through the predictions. Custom solutions were created to, for example, measure distance and area size, upload and visualize shapefiles and export a geospatial pdf to be used during field visits with no internet connection.

Deloitte Impact Foundation

One of the initiatives that make working on a project like this possible at Deloitte is the Deloitte Impact Foundation. This initiative allows us to use our professional expertise to make an impact in society. I am very grateful that this initiative is in place and gives us to the opportunity to contribute to preventing illegal deforestation and protecting the world’s rainforests.”

Did you find this useful?