Workshop 

Leveraging Open-Source Python Tools for Predictive Modeling in Aquatic Resource Management

Organizer(s):

Tropical aquatic ecosystems—from vital fisheries to unique freshwater habitats—are at the forefront of the global conservation crisis. They face mounting pressures from habitat destruction, localized pollution, and the accelerating impacts of climate change. The background to this workshop is the urgent necessity for conservation professionals to move beyond reactive observation and embrace proactive, evidence-based management. This requires leveraging the vast amounts of data being collected into sophisticated forecasts about where risk is highest and where conservation intervention will be most effective. To achieve the conference theme of "Long-Term Resilience," we must democratize access to the tools that make data-driven decisions possible. The scope of this intensive two-hour workshop is to provide participants with a complete, start-to-finish pipeline for predictive modeling using accessible, open-source computational tools. This entire session is built on Python, utilizing Google Colaboratory (Colab), which requires no local software installation, ensuring immediate and seamless participation regardless of technical background. The learning journey is divided into three practical modules: We begin with Geospatial Data Integration, where we use the specialized library GeoPandas to combine tabular environmental data (like water quality) with location coordinates to create compelling, professional thematic maps that visualize conservation hot spots. Next, we delve into the crucial step of preparing this data, often called "Feature Engineering," to structure it correctly for forecasting. The central goal of the workshop is to empower attendees to build their own statistically robust predictive models. Using the popular machine learning library Scikit-learn, participants will train a model (such as a Logistic Regression) designed to predict conservation outcomes, like whether a specific aquatic zone is at High Risk of pollution or decline based on its current environmental variables. By the conclusion, the clear objectives will be met: attendees will know how to manage spatial and tabular data concurrently, execute a predictive model, interpret the model’s results (understanding which factors drive the risk), and ultimately, translate those data insights into practical, actionable policy recommendations for aquatic resource management. This workshop provides the fundamental, reproducible skills needed to lead the next generation of data-driven conservation efforts.