Data science on the world stage

 

September 21, 2014

Work by CEE faculty Bielak, Garrett, and Noh, together with ECE faculty Kovacevic, and University of Pittsburgh faculty Rizzo, was featured at the World Economic Forum's Annual Meeting of the New Champions 2014, in Tianjin, China.


In a world deluged with data and increasingly sophisticated methods for collecting and analyzing it, scientists are focused on how they can use all of this information to improve our lives.

Four Carnegie Mellon University faculty members are providing their perspectives on this topic during a session titled "Data Science in Action" at the World Economic Forum's Annual Meeting of the New Champions 2014, where 1,600 participants from 90 countries are gathering Sept. 10-12, in Tianjin, China.

The Forum's members include the top 1,000 companies from around the world that drive the global economy and collaborate on shaping global, regional and industry agendas.

In 2010, Carnegie Mellon was invited to join the Forum's Global University Leaders Forum (GULF), a group of 25 elite global universities, only 11 of which are in the U.S. Since then, 15 faculty members from six schools and colleges—plus the Entertainment Technology Center—have presented at Forum meetings that attract leaders from business, government, academia and arts and cultural organizations.

Introducing the data science discussion will be CMU President Subra Suresh, who serves as the university's GULF representative and is the only university president serving as one of nine "programme mentors" for the meeting.

James H. Garrett Jr., dean of the College of Engineering, is examining how sensors can be used to collect data from infrastructure, such as water and sewer pipes, buildings and bridges, to locate trouble spots before they break down, which could save trillions of dollars in repair costs.

Dean Garrett on Smart Infrastructure

Imagine a water main that knows it is going to break, closes the right valves in the right sequence, schedules a repair crew, and avoids major collateral damage for citizens, saving hundreds of thousands of dollars that would have been spent in an emergency repair.

This is the future of infrastructure maintenance envisioned by Dean Garrett. His presentation focuses on his field of Smart Infrastructure, which will enable this vision, a field that is a blend of our built infrastructure, with many types of networked sensors that collect data over time and space, and sophisticated analytical techniques used to predict and visualize the conditions of that infrastructure. His work is part of the research in the Smart Infrastructure Institute co-led by Burcu Akinci and Bruno Sinopoli.

In other words, smart infrastructure provides the early indicators of trouble that would compel decision-makers to act in a more timely and effective manner.

In the U.S. alone, trillions of dollars have been invested in civil infrastructure, such as roads, bridges, water systems and sewers. The American Society of Civil Engineers finds that much of this infrastructure is in poor shape, and it estimates that it would take over $2 trillion to improve this situation. 

Unfortunately, for much of our civil infrastructure, we don’t yet have the right tools in place to effectively and efficiently monitor that infrastructure. The data we do collect is manually collected and predominantly based on human visual inspection.

The goal is to better manage infrastructure, saving money and energy and reducing or eliminating disruptions to business and the public.

To do this, Garrett suggests we collect sensor-based data, such as loads, strain and vibrations, and then build models of that data by applying sophisticated analytics. The models enable better understanding of the conditions of the infrastructure, including trends in its performance and options for remediation.

However, placing and maintaining sensors in external environments is a challenging endeavor.

As an alternative, one of CMUs research projects involves exploring the use of vehicles as bridge sensors, instead of installing sensors onto the structure. The vehicles carry, power and protect the sensors and relay the collected information to the analytics station. This yields low installation and low maintenance costs, which is important to the application of hundreds of thousands of bridges. Dean Garrett is working on this research with College of Engineering collaborators Jacobo Bielak, Jelena Kovačević, Hae Young Noh, and University of Pittsburgh's Piero Rizzo and their students.

With smarter infrastructure systems, Dean Garrett and his colleagues anticipate more cost effective management, more efficient operation and more reliable facility and infrastructure performance. Dean Garrett is exploring the possibilities of this research, how it could be implemented, as well as the potential implications for business and government. 

See the slide show video of Dean Garrett's presentation

Learn more about the other CMU data science presenters, as well as the 40 Under 40 participants, including Engineering's Ines Azevedo

Story originally published at: www.engineering.cmu.edu.

Related People:

Jelena Kovačević