Advance Data Science, Artificial Intelligence and Geographic Information Systems (GIS) for Environmental Sciences

Summary 

Course coordinator: Elackiya Sithamparanathan 
Content coordinator: Erkinai Derkenbaeva 
Mail to: elackiya.sithamparanathan@wur.nl  or wimek@wur.nl 
Lecturers: Ir. Gabriela Spakman-Tanasescu and Bastiaan van Dalen

General information

Registration deadline 
Early bird registration deadline: 03 August 2026 
Regular registration deadline: 28 August 2026

Registration link: https://fd8.formdesk.com/wageningenuralterra/data_science_2026 

Remark: As this is an advanced PhD course, we are committed to selecting participants who are the best fit for the course content and objectives. After completing the registration process, you will automatically receive an email containing the link to the online assessment form. The selection of participants will be based on the information you provide in this assessment. 
 
Target group 
PhDs, Postdocs, Assistant Professors, Associate Professors 
 
Group size 
 
Minimum: 5 
Maximum: 25 
 
Credit points 
1.5 EC 
 
Self-study hours 
Circa 8 hours (Depending on familiarity with ArcGIS Pro and Python. For those who are familiar enough, self-study is zero) 
 

Course introduction

This course explores state-of-the-art principles and techniques in Data Science, Artificial Intelligence (AI), and GIS as they apply to complex research questions within the Environmental Sciences. Participants will learn to integrate Geospatial Artificial Intelligence (GeoAI) and spatial Data Science within the ArcGIS (Esri) ecosystem to develop innovative new approaches for their research.
Throughout the course, participants will explore the connections between their own research questions and GeoAI, using either their own datasets or representative samples (e.g., remote sensing data). In this training, we leverage the ArcGIS platform’s capabilities as an open and extensible data science environment. In particular, participants will be introduced to—and encouraged to develop—their own PhD-level mathematical models, solutions, or analytical approaches within ArcGIS Pro using ArcGIS Notebooks.
Beyond the platform’s default functionalities for Data Science (such as the ArcGIS Deep Learning frameworks), emphasis will be placed on leveraging open science libraries where necessary. This illustrates the flexibility of ArcGIS for advanced, research oriented workflows and ensures that analytical tools can be tailored to specific scientific needs.

Learning goals

Our goal is to motivate and inspire participants to add a new dimension to their research by leveraging the spatial component of their data. By utilizing the advanced capabilities of the ArcGIS ecosystem—the world’s most comprehensive Geospatial AI platform—students will gain the skills to integrate, visualize, and analyze complex research questions across both space and time.

Course structure

The course is divided into three integrated components: Theory, Self-study and Exercises, and a Final Assessment.

Theory: fundamentals and deep dives

The theoretical foundation begins with an introduction in GeoAI and Remote Sensing (Day 1), followed by a comprehensive overview of Data Engineering and the core principles of Machine Learning (ML) and Deep Learning (DL).
We will explore ML techniques—such as clustering, classification, and prediction. Alongside DL approaches include image segmentation, object detection, pixel classification, text detection, text extraction and text classification. These methods are applied to diverse datasets, including satellite imagery, aerial photography, oriented imagery, and both structured and unstructured text.
The theoretical track culminates in a "Deep Dive" into Artificial Neural Networks (ANNs) and a specialized look at advanced models and their specific applications within the Environmental Sciences. Participants will learn to leverage ArcGIS Pro’s integrated geoprocessing tools and develop or integrate scientific algorithms using the ArcGIS API for Python within ArcGIS Notebooks.

Exercises: practical application

From Day 1 to Day 4, participants apply theory through hands-on labs designed to bridge the gap between abstract algorithms and real-world spatial data. Starting on Day 2, participants have the choice instead of doing the course exercises to begin their individual assessment, using their own data to practice and apply the theoretical frameworks of the daily topic to their own research (see further paragraph 4.3.). 

Individual assignment and assessment 

The individual assignment follows the CRoss Industry Standard Process for Data Mining (CRISP-DM) workflow or similar academic  structure..
On the final day of the course, each participant will give a 10–15 minute presentation based on the selected topic, presenting the preliminary results achieved during the course week. This is an opportunity to share insights and receive immediate feedback on your methodology and findings, to improve the further work. Participants have one month from the final day of the course to finalize their research and author an "Esri-style" technical article using ArcGIS StoryMaps. This StoryMap must be uploaded to the designated course group (the link to the group will follow). Upon the submission of satisfactory work, students will be awarded 1.5 ECTS credits and will receive a certificate of completion.

For further details on expectations and grading criteria, please refer to the Assessment Rubrics.
For inspiration see the case study example provided below build-up according to the CRISP-DM method:

1.  Research understanding
2.  Data understanding
3.  Data preparation
4.  Modeling
5.  Evaluation
6.  Deployment

This  case study is of Sinne van der Veer (participant at this course in 2025) and can be found here. 

Explanation Sinne’s article structure 
•    Research Question: “Convolutional LSTM for global maize yield prediction—Predicting yields with climate data.”
•    Data Understanding: Sinne aimed to improve yield prediction accuracy using complex GeoAI models by combining the spatial context of satellite-derived data with global historical spatiotemporal yield datasets.
•    Modeling: This project utilized one of the most sophisticated GeoAI models for spatiotemporal prediction: Convolutional Long Short-Term Memory (ConvLSTM). The model uses Space-Time Cubes to aggregate the data in 4D. The data is fed to the model in spatiotemporal batches, effectively adding a 5th dimension to the data. Sinne also implemented a cross-validation with a moving window to split the data into training, testing, and validation data sets, ensuring an optimal model generalization.
•    Evaluation: Sinne concluded with: “ConvLSTM-based approaches could become a valuable tool for agricultural predictions, ultimately contributing to food security and climate resilience.” 
•    Deployment: No deployment step. 

Mandatory required knowledge 

Mandatory required knowledge/ Preparation 

a) Basic ArcGIS Pro skills – mandatory  
 
This is an advanced PhD/postdocs course. The students need to have at least basic knowledge of ArcGIS Pro. Below are some links to help students improve their basic skills if it is necessary: 
- free course “Get started with ArcGIS Pro”: https://learn.arcgis.com/en/projects/get-started-with-arcgis-pro/  

b) Basic Python programming language  not mandatory (a nice to have skills) 
- free course “Learn Python in ArcGIS Pro” https://learn.arcgis.com/en/paths/learn-python-in-arcgis-pro/ (important for this course, the first two lessons).  
 
c) Participants need to bring their own laptops. We suggest a RAM/memory of at least 16 GB, preferably 32GB and a sufficient graphics card. See Deep learning frequently asked questions—ArcGIS Pro If. 
For participants who do not have laptops with a powerful processor (as specified above), we will provide alternative exercises. 
  
d) WUR participants should request an ArcGIS Online account with an ArcGIS Pro license at the GeoDesk of WUR if they do not have one yet. 
Non-WUR participants may need to install ArcGIS Pro using the license of their institutions. 
 
e) The participants should have at least ArcGIS Pro 3.0 because of the Text Detection and Optical Character Recognition capabilities. 

Course fee and cancellation conditions

Fee

The course fee includes coffee, tea and lunch on all 5 days, and dinner on day 1 and drinks on day 5. 
The fee does not include accommodation, breakfast and dinner (apart from dinner on day 1). Accommodation is not included in the fee of the course, but there are several possibilities in Wageningen. For information on B&B’s and hotels in Wageningen please visit proefwageningen.nl/overnachten . Another option is Short Stay Wageningen. Furthermore, Airbnb offers several rooms in the area. Note that besides the restaurants in Wageningen, there are also options to have dinner at Wageningen Campus. 

Cancellation conditions 
•    Up to 4 weeks before the start of the course, cancellation is free of charge. 
•    Up to 2 weeks before the start of the course, a fee of 50% of the full costs will be charged. 
•    In case of cancellation within 2 weeks before the start of the course, a fee of 100% of the full costs will be charged. 
•    If you do not show at all, a fee of 100% of the full costs and a fine of 100 EUR will be charged. 

Course daily structure

Additional information about the Individual Research Assignment (course assignment)

Each student will conduct a small research project. The research question must be directly related to the participant’s own research topic, and the methodology must incorporate at least one of the Machine Learning or Deep Learning models presented during the course.
Deliverables & Format The results of this project will be presented as an ArcGIS StoryMap (the Esri standard for digital articles). See here an example: Detecting and quantifying lunar craters

Evaluation of the StoryMap will be based on the specific criteria.