Namespace:    words-traffic

PI: Volkan Vural
Institution: University of California, San Diego
Project description:

A large traffic dataset joined with environmental data is explored to predict the changes in traffic patterns caused by environmental factors and events such as Padres games, festivals, farmers markets, marathons, sports events and etc. Event based traffic prediction is not only an improvement to existing commercial products but also crucial for many applications including evacuation planning and routing optimization. Due to the nature of the project, large data sets coming from multiple data sources are joined and analyzed using high-performance computing resources at CHASE-CI. Machine learning techniques are used to analyze and predict the traffic with respect to various events. This project is ongoing and currently being extended to the following fields:

  1. Estimating traffic conditions and planning evacuation routes during/after wildfires,
  2. Joining weather station data to predict the impacts of different weather conditions on traffic,
  3. Joining California Highway Patrol data to predict the traffic patterns after an accident, Analysis of how traffic patterns impact the housing market in San Diego County.
Software: Python, Conda, TensorFlow

Back to namespaces