Go Back

Course

MaCAD 23/24 Graph Machine Learning

This course is designed to provide a comprehensive understanding of graph machine learning, from fundamental concepts to some advanced techniques, and will equip you with the necessary knowledge and skills to effectively leverage graph-based approaches in many aspects of design.

Syllabus

Graph Machine Learning 

MaCAD Digital Tools for GRAPH MACHINE LEARNING SEMINAR


Ph credits: Toutable – Ray Harli and Lora Fami – GraML 2022-23

 

In today’s data-driven world, graphs have emerged as a powerful tool for modelling complex relationships and data structures. Graph machine learning, also known as graph-based machine learning or graph analytics, is a rapidly growing field that leverages graph theory, network analysis, and machine learning techniques to extract valuable insights and make predictions from graph-structured data.

This course is designed to provide you with a comprehensive understanding of graph machine learning, from fundamental concepts to some advanced techniques, and will equip you with the necessary knowledge and skills to effectively leverage graph-based approaches in many aspects of design.

Throughout this course, we will cover a wide range of topics, including graph representation, graph embedding, graph clustering, graph neural networks (GNNs), among others. We will also discuss various applications of graph machine learning within the urban domain. 

Through a combination of lectures, hands-on coding exercises, and world case studies we will aim to solidify your understanding of these concepts and techniques. You will also have the opportunity to work with popular graph machine learning libraries and tools, such as NetworkX, DGL and Tensorflow in order to gain practical experience in implementing graph machine learning models and solving design problems.

By the end of this course, you will have a solid foundation in graph machine learning and be able to apply these techniques to analyse and model complex data with graph structure in the urban context. We hope you will enjoy this exciting journey to unlock the power of graphs and take your machine learning skills to new heights!

 

Learning Objectives 

  • Learn the fundamentals of Graph Theory
  • Learn the fundamentals of Graph Machine Learning
  • Learn about the urban networks and space syntax
  • Explore different applications of GraphML within the urban domain
 

Faculty

David Andres Leon

Erida Bendo

Projects from this course

Urban Insights: Leveraging Open Data for Smart Property Rate Predictions

by
MAA 2023-2024

1.0 Concept In the evolving landscape of urban development, property rates are influenced by numerous factors, including geographical, socio-economic, and infrastructural elements. Our project, “Urban Insights: Leveraging Open Data for Smart Property Rate Predictions,” aims to create an intelligent, data-driven model to forecast property rates. This model utilizes features such as wards, property types, and … Read more

Vehicle Crash Prediction in New York City

by
MaCAD 2023-2024

Project Aim In this project we aim to predict the type of vehicle crash that can be foreseen in the city of New York based on the traffic volume, using Graph Neural Networks (GNNs). To develop this machine learning model we use 3 different datasets. The model could hold potential if developed further, to be … Read more

Hotel Prediction for Singapore

by
MaCAD 2023-2024

Our project derives from our observation that in Singapore, most of hotels are located along the east and southeast areas near Changi Airport. There are very few hotels on the west side of Singapore. We also take into account on factors that tourists and visitors consider when making a reservation. These factors include public transportation, … Read more

Re-Routing Bikeways

by
MaCAD 2023-2024

In Switzerland, biking to work is highly encouraged and widely practiced. This is facilitated by the country’s extensive network of bike lanes, bike-friendly infrastructure, and commitment to sustainability However, in terms of safety, Switzerland saw 5,287 cyclist fatalities in 2022, up from 3,793 in 2020. Over the past five years, cycling injuries have increased by … Read more

To bike or not to bike?

by
MaCAD 2023-2024

Bike route classification using Graph Machine Learning The goal of the project was to develop a graph machine learning model that would predict existence of bike routes in Singapore based on collected geodata. Because the bike routes were represented by graph edges, that was also our classification type. Data Exploration & Research Topic We choose … Read more

Phoenix

by
MaCAD 2023-2024

Predicting median income through Traffic Volume, Green and Paved Surface Analysis For our research, we selected Phoenix, Arizona due to its significant disparity in the distribution of green spaces. As depicted in the image, areas with higher income levels display dense green coverage, whereas lower income areas are predominantly paved. The image above depicts the … Read more

Rental Price Predictor – Amsterdam

by
MaCAD 2023-2024

Introduction Accurate prediction of rental prices poses a significant challenge in dynamic real estate markets such as Amsterdam. Our research project explores the use of graph-based machine learning to improve the accuracy of such predictions. This methodology could be of interest to various actors in the real estate sector, including brokers, investors, and urban planners. … Read more

Spatial Analysis of Airbnb Real Estate

by
MaCAD 2023-2024

Our goal was to predict the relationship between the tourist activity zones and the airbnb rentals. Tourism is vital to Spain’s economic growth, with Barcelona as a key contributor, accounting for over 12% of the country’s GDP. TOURIST ANALYSIS AND DATA Tourist spots and zones can be broadly classified to tourist amenities – transits, bus … Read more

by
MaCAD 2023-2024

Dataset Our database contains more than 181,000 rows, each with comprehensive information. The primary database includes 17 variables, though not all are necessary for our analysis. The most crucial data points are location (latitude and longitude), type of food establishment, type of inspection, inspection results, and risk level. As the person who uploaded the database … Read more

MY PARKS : Predicting Miami City’s Parks Scores based on Amenities and Businesses

by
MaCAD 2023-2024
Miami Parks Prediction GraphML Project

Rethinking Urban Spaces Parks and green areas are critical in cities as they provide spaces for people to meet, interact, and find a social life. They contribute significantly to the mental and physical well-being of residents, offering a natural respite from the urban hustle. Project Summary: According to google reviews, the most important factor for … Read more

Graph Machine Learning – A Study on Preventive Strategies for Femicides in Mexico City

by
MAA 2023-2024

For the last two decades, Mexico has grappled with an escalating wave of violence that has put the entire nation on edge, but it is particularly alarming and dangerous simply for being a woman.  This alarming trend, characterized by uncontrolled levels of violence and fueled by a deeply ingrained misogynistic culture, underscores the urgent need … Read more

Project Sentinel: Predicative analysis of street lighting and safety

by
MaCAD 2023-2024

Greater Manchester, one of the largest and most vibrant urban centers in the UK, is characterized by its substantial student population and a dynamic economic landscape. With approximately 120,670 university students during the 2021/22 academic year and a significant number of these individuals studying at the University of Manchester and Manchester Metropolitan University, the region … Read more

AGRO-Graph

by
MaCAD 2023-2024

is a Graph ML Project, to Predict the Potential of agricultural lands according to Soil meniral statistics, River water statistics and waste land types and proximity. Location: Fermanagh and Omagh, Northern Ireland. After extensive research on open-source data, we discovered that Ireland provides a wealth of information. We studied the historical land uses we found … Read more

Predict Yelp Ratings based on Urban Data using GML

by
MaCAD 2023-2024

Hypothesis The goal of this research was to investigate if open spatial data could predict Yelp ratings utilizing graph machine learning (GML) methods. We hypothesize that urban phenomena, events, and objects will indicate customer reviews and popularity, and therefore, could be used predict ratings. In particular, we perform edge classification using the DGL library. For … Read more

Metro Station Prediction in Stockholm

by
MaCAD 2023-2024

This post describes the development of a method to predict metro station locations using Graph Neural Networks (GNNs). Our journey began with a challenge familiar to urban planners: how to strategically place new metro stations to optimize transportation networks. In this case the city of Stockholm was used as testbed. The Challenge of Data Imbalance: … Read more

100-Year Flood Risk Road Intersection Classification

by
MaCAD 2023-2024
Flood Risk Classification

Beginnings Our project’s concept is to develop a classification model that identifies street intersections (graph nodes) in Stockholm, Sweden, susceptible to flooding during a 100-year flood event. Based on our initial research, we found that graph machine learning operates at three levels: the graph, its edges, and its nodes. With access to a 100-year flood … Read more