Long-Term Traffic Volume Prediction for Vancouver

0.0/5
We try to predict long-term traffic volume in hope that it will aid the long-term planning of the City of Vancouver by utilizing machine learning methodologies. We employ a number of machine learning models and time series models such as double exponential smoothing to make long-term predictions based on intersection traffic movement counts data.

School:

Northeastern University

Course:

Computer Science Capstone (CS 79808)

Instructors:

  • Michal Aibin

City of Vancouver:

  • Zak Zenasni

Student Team:

  • Kevin Thomas Abraham
  • Thitikorn Pongmorrakot
  • Cheng Shen
  • Jun Wang
  • Wei Xin

Term:

Fall 2022

Summary

 

A well-designed transportation system is what makes a city an accessible place for everyone, and roadways are an integral part of any city’s transportation network. Regardless of how each of us travels day to day, how well the roads serve their functions is going to have a big influence on our quality of life. Naturally, roadways also play a crucial role in the Transportation 2040 Plan for the City of Vancouver, which is why we would like to be able to predict the long-term traffic volume in order to potentially help us be better informed when making decisions relating to the transportation network. We believe that being able to estimate the future trend in how roads are going to be used is going to be immensely helpful when specific urban planning decisions have to be made, such as in Transportation 2040 or other infrastructure projects that the city of Vancouver might have.

We try to predict long-term traffic volume in hope that it will aid the long-term planning of the City of Vancouver by utilizing machine learning methodologies. We employ a number of machine learning models and time series models such as double exponential smoothing to make long-term predictions based on intersection traffic movement counts data.

We formulated a two-step process: in the first step, as economic factors normally have different impacts on different regions, we divide the city of Vancouver into six regions: Vancouver Northwest, Downtown, South Vancouver, East Vancouver, Vancouver Southwest, and Vancouver Central. We use a Vector Autoregressive model to study the interaction of regional traffic volume and economic factors. In the second step, we employ a number of machine learning models and time series models to make predictions for individual locations (i.e., intersections) by making use of the output from the first step in order to help improve the accuracy of the prediction.

{{ reviewsTotal }}{{ options.labels.singularReviewCountLabel }}
{{ reviewsTotal }}{{ options.labels.pluralReviewCountLabel }}
{{ options.labels.newReviewButton }}
{{ userData.canReview.message }}

Related projects

Stepping Up: Enhancing Urban Mobility, Health, and Inclusion Through Sidewalk Investment in the City of Vancouver

0.0/5

"Functional, well-maintained sidewalks are a crucial component of our communities. In Vancouver, the...

PDS Data Optimization by AI

0.0/5

Our application automates by-law and regulation amendments by downloading relevant files, indexing them...

The State of the Marine Foreshore Habitat of The Wild Mile

0.0/5

We conducted a biodiversity survey on a section of The Wild Mile beach...