Course pre-requisite(s):Basic programming experience in Python
Course Overview
This summer course explores the intersection of data analytics, artificial intelligence (AI), and sustainability. Participants will learn to analyze sustainability data, design and prototype AI-driven solutions, and critically evaluate their impact. Through interactive lectures, hands-on coding sessions, and a culminating group project, students will address real-world sustainability challenges using predictive analytics and AI techniques. The course emphasizes practical application, ethical considerations, and effective communication of data-driven insights for a more sustainable future.
Learning Outcomes
After completing this course participants will be able to:
● Identify and evaluate the impact of data-driven analytics and artificial intelligence technologies in sustainability contexts considering ethical and societal implications.
● Formulate real-world sustainability problems as AI/machine learning tasks, selecting appropriate algorithms and data preprocessing techniques to develop prototype solutions for complex datasets.
● Apply machine and deep learning models to analyze sustainability data (e.g., energy consumption, land use), interpret model outputs, and make predictions to inform decision-making in critical sustainability domains.
Course Content
Week 1: Foundations of Sustainability and Data Analysis
● Introduction to SDGs, global sustainability challenges and related data sources.
● Data collection methods and technologies (remote sensing, IoT sensors, citizen science).
● Data cleaning, preprocessing, and exploratory data analysis.
● Data visualization and communication for effective decision-making.
Week 1: Predictive Analytics for Sustainable Decision Making
● Framing sustainability challenges as predictive tasks and the machine learning paradigm.
● Model development, training, validation and interpretable metrics for impact evaluation.
● Regression and classification techniques using decision trees-based models and ensemble methods.
● Case studies of applying analytics to real-world sustainability problems: energy demand.
Week 2: Fundamentals of Artificial Intelligence for Sustainability
● Introduction to AI concepts relevant to sustainability ( deep learning, computer vision, natural language processing).
● Regression and classification tasks for complex data using deep learning.
● Neural network architectures for sequential and multimedia data.
● Ethical considerations and responsible AI development in the context of sustainability.
Week 2: AI Applications in Sustainability Domains
● AI for climate change mitigation and adaptation: energy optimization.
● AI for sustainable agriculture and resource management: precision farming.
● AI for social and economic sustainability: smart cities.
● Emerging trends in AI for sustainability: limitations and challenges, low-resource data.
This course will not cover the following related topics: in-depth policy analysis, detailed economic modeling, data management, sensor specifics, mathematical proofs of algorithms, in-depth coding implementation of algorithms, advanced topics in neural network architecture design.
Instructional Method
This course employs a blended learning approach combining interactive lectures, hands-on coding sessions, group discussions, and a culminating group project.
● Interactive Lectures: Lectures will introduce key concepts and be interwoven with live coding demonstrations using pre-prepared code snippets, encouraging active student participation.
● Hands-on Coding Sessions: Students will work collaboratively in groups using provided datasets and Jupyter Notebooks, focusing on completing targeted code snippets to analyze data, interpret results, and evaluate impact. This approach prioritizes data interaction and insight generation over extensive programming tasks.
● Roundtable Discussions: Following coding sessions, facilitated discussions will encourage students to share insights, interpretations, and critical reflections on the results.
● Group Project (Two Weeks):
○ Week 1: Problem Definition: Students will use a design thinking-inspired approach to identify a sustainability challenge relevant to their region or interests and define a corresponding dataset. A dedicated feedback session with the instructor and group presentations will conclude the week.
○ Week 2: Prototype Development: Groups will apply the course's data analytics and AI techniques to develop a prototype solution. Ongoing instructor feedback will guide their progress.
○ Final Presentation: The course culminates in group presentations showcasing their chosen problem, proposed AI-based solution, and working prototype.
Required Course Materials
Required Readings:
● Tackling Climate Change with Machine Learning (freely available from CCAI): [ https://doi.org/10.1145/3485128 ] by David Rolnick et al. (2022) provides a comprehensive overview of AI and data-driven applications in climate change as well as the connection with sustainability goals.
● Towards Delivering on the Sustainable Development Goals using Earth Observations (freely available from GEO): [ https://doi.org/10.1016/j.rse.2020.111930 ] by Argyro Kavvada et al. (2020) explores the use of remote sensing data for addressing sustainability challenges.
Software Requirements:
● Laptop or personal computer: with internet access.
● Free Google Account: for access to Google Colab platform (recommended) or local installation of the following tools (instructor will assist with setup): Python, Jupyter Notebooks, PyTorch, and sklearn.
Assessment
Student performance will be evaluated through a combination of group projects, in-class participation, and an individual assessment:
● Group Project (Part 1: Problem Definition & Framing - 20%): At the end of the first week, groups will submit a 2-page project summary document outlining their chosen sustainability challenge and its framing within an AI/data context, followed by a 5-minute pitch presentation.
● Group Project (Part 2: Prototype & Impact - 20%): At the end of the course, groups will submit a 2-page project summary documenting their data/AI prototype, its functionality, and expected impact, followed by a 5-minute pitch presentation.
● In-Class Participation & Coding Challenges (20%): Throughout the course, students will have multiple opportunities to actively participate in discussions and complete short, in-class coding challenges, receiving personalized feedback.
● Individual Assessment (AI & Sustainability Concepts - 40%): A final online test (e.g., using Google Forms) will assess individual understanding of core AI concepts and their application to framing sustainability challenges.