Posted by Datawatch on November 2, 2016

In less than a week, Americans will choose the 45th President of the United States. The latest electoral map still favors Clinton, however Trump could still win the popular vote. Recently, a winning team of college students combined numerous, disparate data sources and performed automated, predictive analysis to make a compelling case detailing how he could do it.

This analysis took place at the Datawatch and IBM Watson Analytics Hackathon, at the University of New Hampshire (UNH) Peter T. Paul Entrepreneurship Center (ECenter) on October 21 and 22. I had the distinct honor of being a judge at the event and was able to witness these students utilize the Datawatch Monarch data preparation platform and IBM Watson Analytics software to make their case.

First, all students received a 30-minute introduction to Datawatch Monarch and a 120-minute demonstration of Watson Analytics. At 4 p.m. ET on Friday, October 21, all 10 student teams were provided with several data sets and the necessary software tools to perform their analysis. Once the teams formulated their hypotheses, they used Datawatch Monarch to unlock and blend data from numerous data sources and formats such as PDFs, CSV files, Excel and Access databases, web content from several published sources and sentiment data from social networks. The prepared data was then processed in IBM Watson Analytics in the cloud, allowing the teams to create data visualizations and dashboards in minutes.

By 10 a.m. ET Saturday, October 22, all teams were required to submit PowerPoint presentations of their analyses. That’s when my fellow judges Andy Smith, director of the UNH Survey Center and Laura Trouvais, academic program administrator of IBM, and I went to work. We evaluated the presentations based on six criteria, including: proficiency in using each tool; creativity and logic in how the analysis was conducted and insights were identified; the usefulness of those insights; and the data visualizations, logic and flow of the presentation.

Ultimately, we selected the winning team, comprised of undergraduate students Brandon Allen, TJ Evarts, Max Miller and Sam Warach. Together, they analyzed U.S. Census data and state polling information, as well as data from the 2012 presidential election to determine the total number of current voters for Donald Trump and Hillary Clinton. Using IBM Watson, they generated a line graph of voter loyalty for each candidate throughout the past 10 months, which revealed that Trump’s core voter base has remained more consistent than Clinton’s. The team determined that if voters cast their ballots “today,” Clinton would win the popular vote by only four percent; however, if Clinton’s voters, who have been historically quick to change their opinion of the democratic candidate, move to a third party, Trump can conceivably win the popular vote.

A glimpse into the winning presentation, “How Donald J. Trump Can Win the Popular Vote.”

A glimpse into the winning presentation, “How Donald J. Trump Can Win the Popular Vote.”

The winning team of the Datawatch and IBM Watson Analytics Hackathon pose with members of the judging committee.

The winning team of the Datawatch and IBM Watson Analytics Hackathon pose with members of the judging committee.

This analysis not only gave these students bragging rights with their hackathon win, they’ve recently returned from an all-expenses paid trip to IBM’s World of Watson conference in Las Vegas that took place last week. There, they had the opportunity to participate in an IBM academic program and present their findings.

While we’ll need to wait until November 8 to find out who wins our upcoming Presidential election, we each feel we’ve walked away as winners from this hackathon experience. Witnessing these individuals – who had no previous experience or proficiency in the blending or analytics tools and just 24 hours – derive value from their data using our data prep software was inspiring and a true testament to the technology itself.

The winning presentation from the Datawatch and IBM Watson Analytics Hackathon was featured at IBM World of Watson, held in Las Vegas, Nevada.

The winning presentation from the Datawatch and IBM Watson Analytics Hackathon was featured at IBM World of Watson, held in Las Vegas, Nevada.

ECenter Program Manager and Hackathon organizer, Heather MacNeill also recently commented, “The ECenter was excited to host UNH’s first Hackathon. One of our priorities is get students from all colleges to interact together around ideas, innovation, and entrepreneurship. The Datawatch and IBM Watson Analytics Hackathon was hugely successful in that effort and students had the opportunity to engage on an interdisciplinary level which was exciting to see.”

 

 

 

See for yourself what award-winning self-service data prep can do for your business. Download your free trial of Monarch today.

 

About the Peter T. Paul Entrepreneurship Center 

The Peter T. Paul Entrepreneurship Center (ECenter) in the UNHInnovation wing of Madbury Commons is intentionally independent of any one college on campus. We are here for all students, faculty, staff, researchers, and alumni. The ECenter’s goal is to continue building the positive and supportive idea/innovation/entrepreneurial culture. We create various opportunities for engagement through new co-curricular programming, as well as supporting existing academic, student, university, and independent initiatives.

We support any stage of idea and start-up development–any discipline or industry, from technology, to social entrepreneurship, to consumer products/lifestyle, and everything in between! Like most university entrepreneurial centers, the ECenter’s focus is not solely on helping to support the creation of start-up companies.  We also focus on helping individuals and teams to understand the process to see problems and find a range of possible ideas and solutions. That learning experience is invaluable to everyone in any career.

 

Product Tour