ɫɫ��

Big Data Analytics to Aid Developing Livable Communities

• PIs: Yang, Cho, Oh – ɫɫ�� Michigan University
• Project Period: July 1, 2014 – August 1, 2015 (12 months)

Big data analytics becomes an upcoming movement in transportation. For researchers, big data offers technical challenges along three dimensions: 1. large data size (such that a data set is too large to be loaded into memory); 2. data streams (the data must be processed “on-the-fly” and you have only one chance to look at it); 3. multiple data types and sources (which demand for integrated and multimodal processing). These challenges are fundamental and invalidate our basic assumptions on traditional data processing. For transportation agencies and users, big data analytics offers opportunities to integrate intelligence into transportation infrastructure, to improve capacity, and to enhance travel experience in a livable community.  This project is proposed to conduct both fundamental research and application development on big data processing in transportation. In fundamental research, we will selectively investigate a few basic problems. Example problems include dimensionality reduction of large data sets, frequency counts of data streams, and Bayesian network model for traffic prediction.  In application development, we propose to setup a working test bed for developing and testing research ideas. The test bed includes a data warehouse, the corresponding web services to answer user queries, and utility functions to collect, integrate, extract and store data from real-time data streams. The test bed also supports data visualization with web interfaces, which allow users to interactively query the data and make use of advanced application tools. Finally, we propose to develop prototype tools for transportation data analysis and decision support. These tools can: 1. analyze both historical data and real-time data streams to identify traffic patterns and to predict traffic conditions in order to reduce road congestion; 2. provide services for personal travel, for example, trip planning by estimating travel time.  The overall objective of this project is to develop techniques and tools for big data analytics in transportation. We hope our effort will be helpful in setting up a primitive stage towards a rigorous framework for general analytical processing of big data in transportation. We hope this effort can eventually contribute to developing livable communities by providing safe and reliable travel services and useful information, such as detailed operational recommendations, to transportation agencies.