During first period at Francis Polytechnic High School in Sun Valley, Monica Casillas asked her students to line up in order by height so she could organize a human representation of a “box-and-whisker” plot.
As they filed into place — most boys went to one end, most girls to the other — Casillas drew the data visualization on butcher paper. The rectangle in the center showed the typical heights of the class, with straight lines called “whiskers” extended from the box to show how far away the tallest and shortest students were from the middle.
Thomas Navas, a 5-foot-2 junior, found himself at the end of a whisker, the shortest student in his Introduction to Data Science class.
Standing at the end, Navas wondered what affected his height. Could his lack of sleep — about six hours a night — keep him from growing?
Asking questions of data is the aim of the class, which is being offered at 10 Los Angeles Unified School District high schools this year. The class gives students an alternative to traditional math; its curriculum is grounded in hands-on data collection, plus lessons in computer programming so students can get answers from data, a trade highly valued in many industries.
The National Science Foundation awarded a $12.5-million grant in 2010 to a partnership including L.A. Unified and the departments of computer science and education at UCLA, in an effort to teach computational thinking in urban schools.
Suyen Moncada-Machado, an instructional specialist with the district, saw a need to use some of the funds for a class devoted specifically to data science. She proposed the course, hoping to accomplish the goals of the grant while also satisfying the district’s recently adopted Common Core guidelines. The new learning standards in math emphasize statistical literacy, specifically probability and modeling — the practice of using data to make informed decisions.
“The current statistics classes were made in the 1970s; we have these powerful computers that do so much more now,” said Moncada-Machado, who previously taught math in L.A. Unified high schools. “If you look at a stats textbook from the 70s and a stats textbook now, there’s not much of a change. We needed something that would bring it into the 21st century.”
Meanwhile, in Casillas’ classroom, Navas dissected a file from the Centers for Disease Control with responses about daily life from thousands of young adults, a data set often used in his class. Using a few lines of code, Navas plotted the relationship between hours of sleep per night and the height of each respondent. His theory held up.
“It’s a different type of approach to math,” Navas said. “Not too many people at school have done it. It’s pretty fun.”
The class has made an impression on students for whom math had not been a strong point. One year after failing Algebra II, Navas was getting an A in Introduction to Data Science, he said.
Others enrolled because it was the only math class that fit into their schedule.
“I didn’t have a clue what [the class] was,” said Jose Bautista, an 11th-grader who couldn’t find an open slot for pre-calculus. “Now I’m into it.”
Bautista said he wasn’t planning to take another math class when the school year started. Now, he’s hoping to take Advanced Placement Statistics during his senior year.
For Casillas, it takes patience to teach a brand-new class. Unlike a long-established course such as AP Statistics, which Casillas also teaches at the high school, her new curriculum will never be the same twice. The AP class is heavy on teaching statistical theory with a steady diet of textbook exercises. Introduction to Data Science relies on surveys students fill out at school and home — on cellphones if they own one — and the responses shape how the class is taught.
“The lessons are written in a way that they’re not lectures so the students can experience it themselves,” Moncada-Machado said. “When they actually get to the abstract stuff they have something they remember and hold on to. It’s experience-based education — doing rather than sitting in a classroom.”
One day, Casillas will ask students to come up with a survey question and ask it of the entire class. The next, students will be learning snippets of a programming language that produce tables and charts in RStudio, a free interface developed by academics for college classes.
In one exercise, students asked each other whether they preferred taking showers in the morning or at night.
Agustin Mendez broke down the answers by gender and noticed girls in the class more often picked the morning and boys preferred showering at night. It made sense to Mendez.
“Girls don’t want their hair messy,” said Mendez, who prefers night showers. His deskmate, fellow 11th-grader Jasmine Villa, laughed in agreement. She likes to shower in the morning.
“We always ask: Why is the data like that?” Mendez said. “There’s always an answer.”
The students at Poly, many taking Introduction to Data Science to replace a math class they failed, are passing, Casillas said.
“We want to make sure they’re passing because they’re learning and not because it’s too easy,” said Rob Gould, vice-chair of undergraduate studies at the UCLA statistics department, who is overseeing the curriculum.
How will they know the class is rigorous enough? This summer, the educators will compile all the projects and tests done by students this year, and analyze the data.