TEXAS OnRamps computer science course engages students with hands-on data scienceby Amanda Voeller
Homicides go hand in hand with happiness in many countries, and height doesn’t necessarily make a basketball player great. TEXAS OnRamps students drew these conclusions and many others through a project in their University of Texas at Austin computer science course, Thriving in Our Digital World, this spring.
OnRamps courses expose high school students throughout Texas to academic experiences aligned with the expectations of leading research universities. Big Data is one of six units in the computer science offering, which provides students with the opportunity to earn three undergraduate credits from UT Austin.
“We are excited about the advanced content knowledge and technology skills students gain through our computer science course,” OnRamps director Dr. Julie Schell said. “The ability to pose big questions about big data using rigorous methods is a valuable skillset applicable in a variety of contexts.”
OnRamps instructor Mr. Erik Dillaman described huge amounts of data, such as the “big data” that Facebook and Google use, as unwieldy and impossible for humans to compute on their own.
“Big data is this collection of huge sets of information from different sources… that [is] so large that humans themselves will never have enough time in their lifetime to make sense of it and see patterns,” Dillaman said. “The computers are able to leverage that. They’re able to see order in what may appear at first to be chaos.”
Although this data may seem unmanageable, OnRamps students across the state are making sense of it by using a data analytics software program called Tableau.
Dillaman introduced Tableau to the course this year. Tableau gives students a way to analyze and visualize large data sets, spot trends and draw conclusions.
Dillaman teaches students to act as data scientists and form hypotheses that the data sets will prove or disprove. In previous years, students used pre-parsed data from Google Public Data Explorer for their projects. However, Dillaman said students had trouble engaging with that data. When students are emotionally and culturally connected to what they are studying, they learn better and put forth more effort, according to Dillaman. So, he decided to allow students to use any data set and ask any question.
“Anything that you want, you can research and study [for this OnRamps project],” Dillaman said. “It’s really tough for a student to say, ‘Oh, I’m not really interested in anything.’”
In Ms. Haley Bolton’s class at Plano Academy, students used Tableau to identify outliers and patterns, illustrating statistically just how much better Kobe Bryant is at playing basketball than the average player. Another group investigated the relationships between hunger and homicide rates. They demonstrated that although hunger and homicide rates aren’t positively correlated, happiness and homicide rates are, a conclusion they didn’t expect.
The OnRamps students work in groups to first learn how to find, interpret and map clusters of data, such as where the most purse-snatching incidents in Austin are and why that might be. They also learn how to illustrate the sizes of different pieces of data. For example, students could create a bubble chart from Austin crime information. Larger bubbles indicate the most common types of crime, such as car theft, while smaller bubbles indicate less common crimes, such as license plate theft. Students then go on to discover on their own by generating questions of data that interest them.
As Dillaman leads students through the project, he hopes they will run into dead ends and re-hypothesize but feel accomplished and ecstatic when they find the answers to their questions. Before this course, many students had never been asked to analyze and draw conclusions from huge sets of data, according to Dillaman.
“They’ve always been given graphs and given visualizations… [and then asked] ‘Alright, well, what does this graph say?’” Dillaman said. “They’re used to being able to deduce information from data that’s already visualized for them, but they’re not used to actually having to create that kind of order out of the numbers… For a lot of them, that was kind of their first foray into that research aspect of academia.”