Prof. Bruno Abrahao: Course Spotlight - Network Analytics

February 24, 2020

Written by NYU Shanghai

Course Spotlight: Network Analytics

Students apply cutting-edge tools and techniques to explore topics ranging from cyberbullying to music industry trends from the perspective of network analytics.

Humans are creatures of networks – we are connected in networks of family members, online communities, and ecosystems. But how exactly are the social, technological, and natural worlds connected, and how do these connections make a difference in business? In Professor Bruno Abrahao’s seven week graduate-level Network Analytics course, students learn how to deal with business problems from the perspective of network analysis and how to use data analytics and programming to conduct the analysis.

Abrahao, who has a background in computer science,  has been using new technologies, such as artificial intelligence and big data, to facilitate business innovation for years. He believes  business students must adopt a complex approach to statistics to compete in the market. “Traditional statistics assumes that [data sets] are independent. However, people connect to each other and influence each other in reality, which means that when you look at two data points, they’re actually one,” says Abrahao. “If you ignore these connections, you're not going to get much value from your data compared to cases where you take these networks into account.”

Abrahao introduces class members to several data analysis tools such as Python, NetworkX, and natural language processing that can be applied to a variety of fields from anthropology to political science.  . For their final projects, students these tools to analyze data in their fields of interest, which ranged from cyberbullying to music industry trends, and product and travel recommendations. 

Phionna Teo, who majored in Anthropology during her undergraduate studies, applied network analytics to the social phenomenon of cyberbullying. “Social media platforms are still a bit slow in using automated systems to identify abusive content. They still rely on user-generated judgements and reporting,” said Teo. “We wanted to identify the types of people who are more susceptible to cyberbullying, so that detection can be automated more efficiently.”
 

Teo’s final project video.

Teo’s team started with a set of 30 thousand tweets from 2013, and trained a classifier to identify “bullying” tweets by using an existing dataset of comments pulled from another study on Wikipedia commenting behaviors that were already labeled “bullying.”  After applying the trained classifier to the data set of 30 thousand tweets, the team ended up with three thousand “bullying tweets” directed at other users. 

Then, they correlated the data set with account information and follower/following counts to see if there were any differences between accounts they had flagged as “bullying” versus unflagged accounts. They also analyzed the data set to find which accounts were “most bullied” on Twitter, and found that Canadian pop star Justin Bieber’s account was the subject of the highest number of bullying and harassing tweets from the 2013 dataset. Finally, Teo’s team used natural language processing tools to analyze word usage, and found correlations between the most commonly used words in bullying comments and the next most highly correlated words to follow them. 

The result of such analyses could potentially be used by a social media platform to automatically detect bullying comments and messages, identify individuals susceptible to bullying, and improve how platforms like Twitter handle online harassment and cyberbullying.

For their final project, Yao Guanhua and Xu Zhixing wanted to see how the topics of songs and their lyrics have changed over time, so they used web scraping to download lyrics from 140 thousand songs from 1970 to 2000, and used natural language processing to identify the topics of the songs. Then they plotted the songs by year and along 15 different topics to show changing trends over time.
 

Yao and Xu’s project tracked changing music trends over four decades.

In the lead-up to their final projects, students study real business cases about network analytics such as a joint study between Airbnb and Stanford University that  Abrahao  participated in when he was a graduate student?  Airbnb wanted to counteract users’ bias toward majority groups. , which showed that their customers tended to gravitate towards guests and hosts with similar ethnicities, religions, and locations.

First, researchers proved that the phenomenon was indeed occurring by inviting one hundred thousand real users to interact with simulated profiles with varied demographics on race, religion, and geographical location. Then, the researchers created networks based on interaction data, and found that people tended to trust strangers with similar demographics more than those with dissimilar ones. The result of this study was used by Airbnb as impetus to continue improving their reputation system, which allows hosts and guests to rate each other.
 
Abrahao’s course also emphasizes the applications of network analytics in Chinese and Asian markets. Students see these applications first-hand through field trips and four guest lectures from academics, entrepreneurs, and executives in the consulting, food and beverage, Internet, and technology industries. “China is in a very special moment right now....the ecosystem is relatively closed off and doesn't follow the rules of the rest of the world,” said Professor Abrahao, “what we can learn here in China is going to be immediately applicable, especially if you decide to work in China or in Asia.”
 
In his guest lecture, Chen Yang, a professor of computer science from Fudan, showed students how he and his research team used machine learning to detect fraudulent accounts on Dianping and Momo -- China’s Yelp and OkCupid. “Professor Chen’s lecture was very clear and technical,” said Teo. “He showed us how Chinese apps were operated, and we learned that many of the analytical tools from class could be applied to real cases.”

Students also heard from Reinaldo Normand, founder and investor of multiple start-ups in Silicon Valley, who shared his career experience and gave the class entrepreneurial advice. “What I found interesting about his lecture were the differences between Chinese and American corporate structures and entrepreneurs’ personalities,” said Xu Zhixing.

Abrahao also took his students to visit the new Amazon Joint Innovation Center in Shanghai’s Jing’an District, and the Shanghai Data Exchange Center -- a central hub where big data from Yangtze River Delta Region and the entire country is integrated and analyzed for purposes like business, transportation, security, and urban planning.

“We were so excited to learn how real companies are using new technology,” said Yao Guanhua. “At Amazon, we learned how they used cloud computing and big data to build intelligent systems in China. And at the Data Center, they showed us how they stored, maintained, and used massive amounts of data for marketing, entrepreneurship, and innovation.” 

This course is open every fall semester to graduate students in the MS in Data Analytics & Business Computing program.