🧐 Lectures#
Table of Contents! 👇🏽
Tip
A GUIDE TO FOLLOW THIS PAGE
The slides will be updated latest a night before the lecture in pdf format.
Lectures will not be recorded or delivered online.
The section To do before class provides content that is useful for following the lectures. I expect you to follow it before every lecture. It will take about 1 hour of prep at home.
Section Extra Material is exactly extra. It is not required for this course but can prove really helpful for gaining extra knowledge either during or after this course. Sometimes I use it to build the contents of the lecture and at others I find them helpful in my research related to the weekly topics but I will never question your knowledge on that.
Let’s begin#
Before starting the course, watch this video by Khalid Kadir about a reflection on poverty (an example of a social problem), expertise and equity. This representation is an example of how experts create boxes around their craft. As a data scientist or a future expert (consultant, data analyst, policymaker, etc.), it is our responsibility to step out of those boxes and engage with communities to strive for just outcomes. If you have seen this video, send me a meme about poverty, inequality and data. The best submission will win a prize at the end of the course.
Lecture 1 - Spatial Data Science#
To do before class [Takes about 1 hour of prep at home]#
As a way to whet your appetite about the content of the first class, I recommend you:
Listen to this interview with Hilary Mason, Max Shron, and Alex Pentland about the power of data.
Watch this video by Mike Flowers, Chief Analytics Officer, at the City of New York about how data is used to influence policy decisions.
Read What New Yorkers are complaining about and reflect on if the cost of running such data systems worth the price of knowing?
Extra Material [Always to learn more but never needed for the course]#
“Chapter 1: Introduction” (Schutt & O’Neil, 2013). Free sampler of the book containing the chapter available online (html, pdf).
Read this critical argument about objectivity and positionality: How Does Your Positionality Bias Your Epistemology?
A Geographic take on Data Science, proposing a new field called Geographic Data Science
Read this short critical piece called Making Space in Geographical Analysis
Lecture 2 - Spatial and Urban Data#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Watch the TED talk by Carlo Rati about MIT’s SENSEable City Lab projects: excellent set of examples
Read the New York Times piece on US buildings map
Explore the Global Human Settlement Layer Dataset, by the European Commission
Extra Material [Always to learn more but never needed for the course]#
The part of the lecture on new sources of data relies on Arribas-Bel, 2014 and Lazer & Radford, 2017.
A classic on the rise of volunteered geographic information.
Lecture 3 - Data Grammar#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Read a very influential pro-big data/data science article, The End of Theory by Chris Anderson, the editor in chief of Wired.
Read A reflexive call of caution on Big Data Analytics by David Lazer et al.
Read Creating healthy and sustainable cities: what gets measured, gets done .
Extra Material [Always to learn more but never needed for the course]#
A cheatsheet (such a misnomer – nobody is cheating and it is a helpful and beautiful resource) on Data Wrangling with Pandas that you may want to stick to your wall or put as your screensaver to save time on finding useful and operational codes.
Lecture 4 - Data Engineering#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Read a short blog on Why, How and When to Scale your features
Extra Material [Always to learn more but never needed for the course]#
The contents of this lecture are loosely based on, and explored into further detail, in the following two references :
Section 9.3.1 of The Hundred-Page Machine Learning Book by Andriy Burkov.
A more academically suited blog on Feature Scaling
Lecture 5 - EDA and Visualisation#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Think about the grammar of graphics when improving your graphs - at Colourful Facts – a Medium blog by Thomas de Beus. Ignore the reference to the R programming language as this course is based on Python (no offence intended to any community, R is the best for visualisation though).
Learn about Kernel Density Estimation
Extra Material [Always to learn more but never needed for the course]#
Berinato, S. Visualisations That Really Work, Harvard Business Review, Jun 2016
Alberto Cairo’s weblog called The Functional Art about information design, and visualisation is an excellent resource for improving your visualisations.
(Yau, 2011)’s book “Visualize this” is a good general introduction to visualisation.
Check out From Data to Vis chart selector for selecting the right charts
Lecture 6 - Geo-Visualisation#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Watch this lecture on “Statistical maps” by Luc Anselin (link to 25min video).
Read the Conversation piece on the Flint case, where the MAUP played a key role.
Spend the rest of the prep hour browsing through Nathan Yau’s excellent blog, Flowing Data.
Extra Material [Always to learn more but never needed for the course]#
(Brewer, 2015)’s Designing Better Maps covers several aspects of building compelling geo visualisations.
Choropleth chapter from the GDS Book (in progress).
Color palettes are important for maps. Find some in ColorBrewer.
Lecture 7 - Networks and Spatial Weights#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Read Eli Knaap’s blog on Measuring Urban Segregation with Spatial Computation
Watch this lecture on “Spatial Weights” by Luc Anselin (link to 34min video). Keep in mind the motivation, in this case, is focused on spatial regression.
Lecture on “Spatial lag” by Luc Anselin (link to video, you can ignore the last five minutes as they are a bit more advanced).
Extra Material [Always to learn more but never needed for the course]#
Check out Geoff Boeing’s computational notebook showcasing the use of OSMNX- a python library for processing street networks as network objects- with a case of Urban Street Network Analysis
For advanced and in-detail treatment, (Anselin & Rey, 2014) is an excellent reference.
Lecture 8 - Exploratory Spatial Data Analysis#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Extra Material [Always to learn more but never needed for the course]#
Reflections on spatial autocorrelation by a quantitative geographer.
Lecture 9 - Machine Learning for Everyone#
You may want to buy The Hundred-Page Machine Learning Book as some chapters will be used in some topics from this point onwards and it is generally a fantastic book to have. If you cannot or do not want to spend $20.00 on the e-copy, email me, and we will figure something out. The author has invested a lot in writing this book, and it is an excellent resource on Machine Learning, even beyond this class.
Slides#
To do before class [Takes about 1 hour of prep at home]#
Extra Material [Always to learn more but never needed for the course]#
The contents of this lecture are loosely based on, and explored into further detail, in the following two references :
Chapter 1 and section 3.1 of The Hundred-Page Machine Learning Book by Andriy Burkov.
Lecture 10 - Anatomy of a Learning Algorithm#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Go through this medium post about A first step toward the future of neighbourhood design
Read this critique about neighbourhood design - A visit to the smart-city-in-progress at Sidewalk Toronto prompts questions about what it means to “participate” in civic design.
Extra Material [Always to learn more but never needed for the course]#
Lecture 11 - Clustering#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Talk on “Geodemographics and the Internal Structure of Cities” by Prof. Alex Singleton (link to 50min. video).
Extra Material [Always to learn more but never needed for the course]#
Supervised Regionalization Methods: A Survey is an excellent review of regionalisation algorithms, but it is an excellent read.
Lecture 12 - Dimensionality Reduction#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Read through this excellent step-wise example of Principal Component Analysis using airport delay data
Read this excellent community-driven explanation of PCA on StackExchange.
Extra Material [Always to learn more but never needed for the course]#
I recommend reading this amazing paper by Caitlin Robinson on A spatial perspective of energy poverty.
If you are feeling adventurous, you can read my work on Inequalities in solar adoption in The Hague.
Lecture 13 - Spatial Density Estimation#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Extra Material [Always to learn more but never needed for the course]#
None…
Lecture 14 - Responsible Data Science#
Slides#
To do before class [Takes about 1 hour of prep at home]#
Read A city is not a computer, Shannon Mattern which carefully examines the limitations of computation in bettering the human condition.
Explore the Gender Shades project by Joy Buolamwini and Timnit Gebru that uncovers the priorities, preferences and prejudices of influential organisations that develop automated systems.
Extra Material [Always to learn more but never needed for the course]#
Watch The unfortunate history of racial bias in photography..
Watch Coded Bias on how corporations are not held accountable for deploying algorithms that affect humans.