Using deep learning and time series forecasting to reduce transit delays

This will be presented in English.

Mark Ryan (IBM), Alina Li Zhang (Skylinerunners)
14:5015:30 Thursday, June 20, 2019

必要预备知识 (Prerequisite Knowledge)

  • A basic to intermediate understanding of machine learning
  • Familiarity with deep learning

您将学到什么 (What you'll learn)

  • Learn to apply time series forecasting in machine learning and apply machine learning to resolve fundamental transit issues

描述 (Description)

Toronto is unique among North American cities for having a legacy streetcar network as an integral part of its transit system. The streetcars are icons of the city and make a unique contribution because of their passenger capacity and zero emissions. However, because most of the streetcar network shares streets with other vehicles, and because unlike buses, streetcars cannot move around obstacles or be quickly towed away if they break down, streetcar delays have a disproportionate effect on gridlock.

Mark Ryan and Alina Li Zhang demonstrate an approach that uses time series forecasting and deep learning to predict streetcar delays. You’ll learn the economic impact of the streetcar-trigged gridlock on Canada’s largest city and the development environment Mark and Alina used to tackle the problem. They outline the source data and cleansing process with a focus on the challenges and opportunities of dealing with wild data and share the essential points of time series focusing as they applied it in conjunction with deep learning. You’ll see the results of their analysis, including practical recommendations for Toronto’s transit authority (the TCC) to resolve the problem of streetcar delays so the city can continue to enjoy the character and soul of its unique street railway for the rest of the twenty-first century.

Photo of Mark Ryan

Mark Ryan


Mark Ryan is a leader in the machine learning hub at IBM, and he’s had an interest in machine learning and artificial intelligence since doing a masters at the University of Toronto in the ‘80s. At IBM he’s responsible for shepherding customers to a variety of database products, including IBM Integrated Analytics System, which includes a full-blown machine learning environment: DSX. Mark’s interests in machine learning include deep learning on structured data and natural language processing (NLP).

Photo of Alina Li Zhang

Alina Li Zhang


Alina Zhang is a data scientist at Skylinerunners and is certified as a Google Cloud Professional Data Engineer. She’s authored articles on machine learning, exploratory data analysis, data visualization, etc. Alina is driving Skylinerunners to provide small businesses with AI solutions. She applies machine learning models on user behavior analysis, recommendation systems, and time series forecasting. Previously, Alina was data scientist in Nobul, where she was driving Nobul to evolve real estate in the cloud with machine learning technology to a variety of problems, including property listing prediction, a real estate chatbot with natural language processing, customer’s behavioral clustering, etc. She was a software developer and WLM component owner of IBM DB2 for IBM. Alina holds a masters degree in computer science from Western University, where her research focused on high-performance computing and truncated Fourier transform.

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)