AIM4Mobility: Predictive Modeling and Demand Forecasting
The future of urban mobility relies on data-driven decisions
Training Format
Self-paced course.
Duration
8 hours (considering theoretical and practical tasks)
Language
English
Certificate of completion
Upon course compeltion, each participant will recieve a diploma from the EIT Urban Mobility.
What will the participants get from the course
Apply
machine learning and AI techniques to real-world urban mobility problems.
Build and evaluate
predictive models using urban mobility datasets.
Understand
spatial clustering and geospatial machine learning methods for identifying mobility patterns and urban analysis.
MODULES
URBAN MOBILITY AND PREDICTIVE MODELING
Introduction
This module introduces the main challenges in urban mobility and how predictive modelling can help address them.
You'll explore the role of bike-sharing systems in modern transport and gain a foundational understanding of predictive modelling and the machine learning process. The module also surveys key ML algorithms used for mobility prediction, focusing on bike-sharing applications.
A final assessment consolidates your learning before moving deeper into modelling techniques.
You'll explore the role of bike-sharing systems in modern transport and gain a foundational understanding of predictive modelling and the machine learning process. The module also surveys key ML algorithms used for mobility prediction, focusing on bike-sharing applications.
A final assessment consolidates your learning before moving deeper into modelling techniques.
DATA COLLECTION AND UNDERSTANDING BIKE-SHARING SYSTEMSModule 1
In this module, you'll explore how bike-sharing systems operate and how data is collected across different touchpoints. You'll get hands-on with common bike-sharing datasets, learning about their structure, key variables, and environmental, temporal, and socio-economic factors that influence usage patterns.
By the end, you’ll have a strong foundation to analyse and model bike-sharing data effectively.
By the end, you’ll have a strong foundation to analyse and model bike-sharing data effectively.
DATA PREPROCESSING FOR MOBILITY PREDICTIONModule 2
This module focuses on preparing high-quality data for accurate mobility prediction.
You’ll learn how to handle missing values in time-series datasets, engineer meaningful features, and apply transformation techniques to capture temporal and spatial patterns. By the end, you’ll understand how robust preprocessing improves model performance and prediction reliability in urban mobility contexts.
You’ll learn how to handle missing values in time-series datasets, engineer meaningful features, and apply transformation techniques to capture temporal and spatial patterns. By the end, you’ll understand how robust preprocessing improves model performance and prediction reliability in urban mobility contexts.
EXPLORATORY DATA ANALYSIS FOR BIKE-SHARING SYSTEMSModule 3
This module introduces key exploratory data analysis (EDA) techniques for understanding bike-sharing systems.
You’ll learn to extract statistical insights, visualize temporal trends, and identify spatial usage patterns. By uncovering hourly, daily, and seasonal behaviours, as well as geographic hotspots, EDA equips you with the foundational knowledge needed to guide deeper analysis and predictive modelling.
You’ll learn to extract statistical insights, visualize temporal trends, and identify spatial usage patterns. By uncovering hourly, daily, and seasonal behaviours, as well as geographic hotspots, EDA equips you with the foundational knowledge needed to guide deeper analysis and predictive modelling.
VISUALIZING URBAN MOBILITY
Module 4
In this module, you'll learn how to create compelling visualizations to communicate mobility trends using interactive mapping tools.
You'll explore the basics of Folium, including how to add markers, polygons, and cluster accident data. You'll also generate choropleth maps to reveal spatial patterns in accident frequency and traffic flow. These visualization techniques will help you better interpret and present mobility insights.
A final assessment wraps up the module and sets the stage for spatial analysis.
You'll explore the basics of Folium, including how to add markers, polygons, and cluster accident data. You'll also generate choropleth maps to reveal spatial patterns in accident frequency and traffic flow. These visualization techniques will help you better interpret and present mobility insights.
A final assessment wraps up the module and sets the stage for spatial analysis.
MACHINE LEARNING MODELS FOR DEMAND FORECASTING Module 5
This module explores the use of machine learning to predict bike-sharing demand.
You’ll implement linear regression for baseline forecasting and expand to decision trees and random forests to capture non-linear patterns. Specialised time series approaches are also introduced, including ARIMA and time-based feature engineering.
The module concludes by comparing model strengths and selecting appropriate techniques for various mobility forecasting scenarios.
You’ll implement linear regression for baseline forecasting and expand to decision trees and random forests to capture non-linear patterns. Specialised time series approaches are also introduced, including ARIMA and time-based feature engineering.
The module concludes by comparing model strengths and selecting appropriate techniques for various mobility forecasting scenarios.
MODEL EVALUATION AND SELECTIONModule 6
This module highlights the critical role of model evaluation and selection in accurate mobility prediction.
You'll explore key evaluation metrics, understand their tradeoffs, and apply time series–specific cross-validation techniques to avoid data leakage. The module also covers model selection and hyperparameter tuning using grid and random search.
By the end, you’ll be equipped with best practices for assessing and refining models to ensure reliable bike sharing demand forecasts.
You'll explore key evaluation metrics, understand their tradeoffs, and apply time series–specific cross-validation techniques to avoid data leakage. The module also covers model selection and hyperparameter tuning using grid and random search.
By the end, you’ll be equipped with best practices for assessing and refining models to ensure reliable bike sharing demand forecasts.
AIM4Mobility is made up for 4 courses
Tools and Techniques for Geospatial ML
This course teaches geospatial data analysis with Python to tackle urban mobility challenges, using OSM data and spatial tools.
Fundamental Trends of AI & ML
The couse covers applications like demand forecasting, traffic optimisation, and autonomous vehicles, as well as key tools, challenges, and ethical issues.
Predictive Modeling and Demand Forecasting
This course teaches how to build and apply predictive models in urban mobility using AI and ML.
AI Challenges & Ethical Considerations
Introduction to ethical principles for AI in urban mobility, covering oversight, accountability, data governance, and fairness.
This project is supported by EIT Urban Mobility, an initiative of the European Institute of Innovation and Technology (EIT), a body of the European Union. EIT Urban Mobility acts to accelerate positive change on mobility to make urban spaces more liveable. Learn more: eiturbanmobility.eu
