Climate Solutions

CSOL-208: Data Science for Climate Solutions

Fall 2026 • 2 Units

Wednesday, 12:00pm - 1:49pm

311 Wellman Hall

This course integrates data science fundamentals into the climate solutions curriculum, reframing them for an AI-driven era. Students practice data organization, visualization, and quantitative analysis by communicating in natural language with AI systems. We explore modern code-developemnt clients, MCP tooling, LLM API use, local models, and development of AI agents. Alongside hands-on, team based projects, we also critically examine AI's reliability, ethics, and environmental impacts.

This course is part of the Master of Climate Solutions program at UC Berkeley, a professional degree designed to train the next generation of climate leaders. The MCS program combines rigorous technical training with practical skills in policy, finance, and leadership to address the climate crisis. Students interested in the program can learn more and apply at climatesolutions.berkeley.edu.

Instructors

Carl Boettiger

Carl Boettiger

Associate Professor, ESPM

Carl works on ecological forecasting and decision making under uncertainty, focusing on regime shifts and data science approaches to environmental problems. Co-founder of rOpenSci and the Schmidt Center for Data Science & Environment at UC Berkeley.

[email protected] Website

Office Hours

Fridays, 12:00 PM

Zoom Link

Gaby May Lagunes

Gaby May Lagunes

Graduate Student Instructor (she/her)

Gaby is a PhD student in ESPM at UC Berkeley, working on climate models, hydrology, and data science. She holds a master's in Data Science and Information from the Berkeley I School and brings 5+ years of industry experience across a range of data roles, along with experience as a D-Lab consultant.

[email protected] D-Lab Profile

Office Hours

Mondays, 4:00–5:00 PM

14 Weeks

One session per week, ~2 hours each

Team-Based

Hands-on collaborative projects

MCS Core

Required for Master of Climate Solutions

The Agent-First Methodology

We do not memorize syntax. Instead, we learn to:

Architect

Design data flows

Prompt

Direct AI implementation

Audit

Verify outputs

Deploy

Ship solutions

Four Core Modules

14 sessions over 14 weeks

Module 1

The AI-Data Analyst

Sessions 1-4

Build an interactive emissions dashboard. Learn AI-assisted coding, data cleaning, visualization, and database querying.


HTML/CSS/JS DuckDB Pandas

Module 2

Spatial Data & Environmental Justice

Sessions 5-7

Map data center environmental impacts. Analyze geospatial patterns, demographic data, and biodiversity impacts through an environmental justice lens.


DuckDB Spatial maplibre Rasterio

Module 3

Working with LLMs & Unstructured Data

Sessions 8-10

Extract structured data from sustainability reports. Work with LLM APIs programmatically, structured outputs, and Model Context Protocol.


LangChain OpenRouter MCP

Module 4

The Capstone Studio

Sessions 11-14

Build a deployable MVP. Scope projects, sprint development, refine user experience, and present live demos.


MVP Demo Day Deployment

Core Tech Stack

Industry-standard tools for rapid prototyping

VS Code / Kilo Code

AI-augmented IDE and coding agent

HTML / CSS / JS

Static web apps & dashboards

DuckDB

Data & spatial queries

LangChain

Document parsing

Questions?

For more information about this course, please reach out to the instructor.