Update, September 2017 : I am no longer a Berkeley GSI, but Paul’s class is ongoing. Check out his teaching materials in his GitHub repo. From my experiences here, I have developed a course series on urban data science with Python and Jupyter, available in this GitHub repo.

The fall semester begins next week at UC Berkeley. For the third year in a row, Paul Waddell and I will be teaching CP255: Urban Informatics and Visualization, and this is my first year as co-lead instructor.

This masters-level course trains students to analyze urban data, develop indicators, conduct spatial analyses, create data visualizations, and build interactive web maps. To do this, we use the Python programming language, open source analysis and visualization tools, and public data.

This course is designed to provide future city planners with a toolkit of technical skills for quantitative problem solving. We don’t require any prior programming experience - we teach this from the ground up - but we do expect prior knowledge of basic statistics and GIS.

Teaching agenda for this semester:

Learning basic Python coding: we spend about 3 weeks introducing the fundamentals of coding, data types, conditionals, loops, and functions. Just enough to make you dangerous.

The scientific computing Python stack, via Anaconda: we introduce numpy, scipy, statsmodels, pandas for data analysis, and matplotlib for visualization.

CartoDB for simple interactive web mapping.

Leaflet and MapBox: more advanced web mapping for community engagement, advocacy, and visualizing research findings.

Installing, customizing, and using Wordpress: each student creates a customized blog as a web portfolio to share and publicize their projects from class.

GitHub for version control and collaborative development.

QGIS: a powerful, free, and open-source alternative to ArcGIS.

Data wrangling, regression, spatial statistics, and spatial analysis with Python and the pandas and geopandas libraries.

Machine learning: including feature engineering and selection, dimensionality reduction, modeling, and clustering in Python with scipy and scikit-learn.

Modeling and visualizing urban accessibility, walkability, street networks and the built form.

Working with APIs, including the census bureau, Socrata (via the City of Berkeley’s open data portal), Google’s geocoding services, and the Twitter API (including mapping and visualizing tweets).

Critical thinking skills about urban data and big data: how to formulate research questions, conceptualize possible solutions, and interrogate problems with equity and representation.

Our guest speakers this semester include Alicia Rouault and Prashant Singh - the CEO and CTO, respectively, of LocalData; Michal Migurski, the CTO of Code for America; Eddie Tejeda, the CEO of Civic Insight; Donna LaSala, the Director of Information Technology for the City of Berkeley; AutoDesk’s Matt Davis; and my fellow doctoral student, Sam Maurer.