Week 6

Statistical Inference for Urban Data

This week, we will discuss using spatial and urban machine learning to analyze and interpret urban phenomena. As we saw last week, urban data has unique features (autocorrelation, heterogeneity,) so we must consider those properties when building our inference, prediction, and descriptive models.

In our practical exercise, we will build on those ideas and the results from our previous lab to analyze the relationship between income and other demographics using spatial regressions and machine learning models.

We also have our first homework released!

Prepare

📖 Read some review papers about Spatial Data Science and Machine Learning:

• Integrating explanation and prediction in computational Social Science by Hofman et al. [1]

• Spatial machine learning by K. Kopczweska [2]

• Becoming a Spatial Data Scientist. E-book by the Carto Company.

• Models for Spatial Data from the Spatial Data Science book by Edzer Pebesma and Roger Bivand

• Statistical Learning from the Geocomputation with R book by Robin Lovelace, Jakub Nowosad, and Jannes Muenchow

Participate

🖥️ Lecture 6 - Inference and prediction for Urban Data

Perform

⌨️ Lab 6-1 - Spatial Inference and Machine Learning

Homework

📝 Homework 1 - 15-minute city in US

Back to course schedule ⏎

References

[1]

J. M. Hofman et al., “Integrating explanation and prediction in computational social science,” Nature, vol. 595, no. 7866, pp. 181–188, Jul. 2021, doi: 10.1038/s41586-021-03659-0.

[2]

K. Kopczewska, “Spatial machine learning: New opportunities for regional science,” The Annals of Regional Science, vol. 68, no. 3, pp. 713–755, Jun. 2022, doi: 10.1007/s00168-021-01101-x.