Smart Devices Drive Demand

The demand for data scientists is partly driven by the increasing number of smart devices in the marketplace for both businesses and private individuals. Common, everyday devices now collect and transmit data—from refrigerators that keep track of expiration dates to traffic signal sensors that adjust brightness levels in response to ambient light.

“We just have a lot more data than we ever did before,” says Geoffrey Parker, director of the Master of Engineering Management Program and professor of engineering. “We’re awash in data, but it doesn’t do much good if it’s just a bunch of zeros and ones.”

As a result, the field of data science has expanded exponentially, and businesses have realized they have to start making use of the data they collect in order to remain competitive.

A 2019 LinkedIn study found that data scientist is the most promising job in the country. According to the U.S. Bureau of Labor Statistics, data scientist is one of the fastest-growing occupations, with a 31-percent projected growth rate from 2019 to 2029.

“We now see tons of organizations grappling with the fact that they’ve got data, but they don’t have the people or the systems to make use of it. They’re trying to staff up,” says Parker. “Data science matters because all that data is just zeroes and ones until you analyze it and turn it into actionable information. It’s one thing to just bang away at a data set, but much more important is asking ‘Why? What am I trying to figure out? What am I trying to solve for?’ ”

NSF Grant Expands the Pipeline

To answer these questions and meet the exploding demand for data scientists, Dartmouth has launched a new initiative to expand the pipeline across STEM disciplines. Thanks to a four-year, $2.8-million grant from the National Science Foundation (NSF), a team of engineering and computer science faculty, along with graduate and undergraduate students, is developing new data science for STEM courses. The Data Science Infused into the Undergraduate STEM Curriculum (DIFUSE) team has been working individually with faculty across the College since the fall of 2019 and has so far developed eight modules that can be integrated into current Dartmouth undergraduate course curricula, such as “Climate Extremes on a Warming Planet,” “Statistical Methods in Engineering,” and “Introductory Psychology.”

The group is led by four principal investigators (PIs), including engineering professors Laura Ray and Petra Bonfert-Taylor; Lorie Loeb, faculty director at the Digital Applied Learning and Innovation (DALI) Lab; and math professor Scott Pauls. An initial team—including one PI, one graduate student, and one undergraduate student—meets with a course instructor to develop learning objectives that help tailor data science modules to fit course-specific needs. The entire DIFUSE team then develops the module in collaboration with the course instructor to ensure it meets the desired learning outcomes.

“We work with instructors of introductory STEM classes who are interested in having a module developed for their class that introduces students to basic data science techniques and practices,” says Bonfert-Taylor. “Each module we have developed so far is completely unique—no two are even close to similar.”

Faculty response has been positive. “I really enjoyed working with the DIFUSE team to get more data science into my course,” says Robert Hawley, chair and associate professor of earth sciences who worked with DIFUSE to develop a module for his “Environmental Change” class. “Using the additional data module from the DIFUSE team, my students were able to extend their analysis to the next level. As a result, nearly every student group connected the ideas that I’d hoped for in the project, where previously many groups still missed important components of the concept.”

Biological science assistant professor Caitlin Hicks Pries agreed her ecology students benefitted from the newly incorporated module. “This data-exploration lab is one that I have been wanting to design, but I did not have the time or technical know-how to do it on my own,” she says. “It was so much fun to work with two DIFUSE students who were so dedicated to the success of this project. From them I learned about translating my code to an interactive webpage, and, I hope, they learned a bit about carbon balance and eddy covariance from me.”

Students Learn New Skills

By infusing data science into the STEM disciplines, DIFUSE seeks to enhance undergraduate students’ expertise in data science and prepare them to enter the workforce. In addition, the dozen or so students on the DIFUSE team have gained experience building effective and engaging courses.

“We want students to feel motivated to learn R [a programming language for statistical computing], and we want this module to fit in seamlessly,” says engineering major Sarah Korb ’22, who helped students in “Statistical Methods in Engineering” use R for statistical analysis. “As a student who is normally on the receiving end, it is impressive to see how much thought goes into building a Dartmouth course.”

DIFUSE students were also able to boost their own learning. James Busch, a PhD student studying earth science, has worked on two modules. For “Stars and the Milky Way,” he completed a series of activities that enabled students to use Python libraries to learn about celestial objects in the galaxy. “I really enjoyed learning about the basics of astronomy and how we view objects in our very own galaxy while designing the exercises,” he says.

An environmental studies module Busch worked on for “Environment and Society” was especially rewarding. After coming across research that linked increased COVID-19 mortality with race and regions of Louisiana, he helped create a web app to examine deaths, air pollution, and demographic data from the state. The app allowed students to investigate the relationships between the Louisiana communities with historic pollution and increased COVID-19 mortality.

“We built the application so students could get hands-on learning with important environmental social justice issues such as those that have occurred in ‘Cancer Alley’ and work with a highly relevant COVID-19 dataset that is currently being studied by researchers of all different fields in the United States,” says Busch.

He enjoyed working on both modules. “I was very impressed by the PIs and the creative freedoms they gave the students in developing the modules—it made the experience that much more fun and gratifying.”

Data Science for All

As DIFUSE was gearing up in the fall of 2019, another data science program sprang up on campus. Dartmouth Engineering launched an online program to offer a professional certificate in data science in partnership with online learning company Emeritus. The course, which is open to all, is based on Parker’s popular “Data Analytics” engineering management class.

Students learn data science fundamentals and high-demand skills such as data visualization, machine learning, risk management, and predictive capabilities, with a focus on real-world concepts. Students work directly with industry mentors and complete a final portfolio demonstrating software, math, and engineering skills.

At the end of the program you actually have the ability to solve practical data analytics problems within an organization,” says Parker. “I think that there’s a certain beauty in being able to visualize data and solve problems and saying, ‘Aha! That’s really neat, and I didn’t know that before.’ ”

New six-month sessions launch every two months, and the only prerequisite is a knowledge of calculus, linear algebra, statistics, and probability. Student can expect to spend 10 to 15 hours each week on the course, which Parker recommends for those who want to pursue a career in data science.

The Next Digital Transformation

Dartmouth Engineering will launch an online program focused on digital transformation, this time in partnership with Coursera. While not exactly data science, participants will learn how to integrate artificial intelligence (AI) and machine-learning solutions to streamline business operations, lower costs, and respond to new market opportunities.

The program will include live-session classes and draw from the expertise of the MEM program at Dartmouth and will be taught by Parker and engineering professors Elizabeth Murnane and Vikrant Vaze.

Participants can expect to spend eight to 10 hours per week on the six-month program and earn a certificate in digital trans-formation. There are no formal prerequisites, but students are encouraged to have an interest in digital transformation and innovation, experience with college-level learning, and some familiarity with Python.

A Pioneering Role

As technologies continue to evolve, so will the field of data science. Parker predicts improved automation will decrease the time and effort required by humans to assemble and clean data sets, leading to a demand for different skills. “I think that technical aspects will get suppressed as they get absorbed into our cloud solutions,” he says. “We could then shift the 70 or 80 percent of effort that’s currently consumed by building data sets over to analyzing them. What will be left is understanding your problem and explaining results in a meaningful way.”

In order to help the emerging workforce keep up with the field, the DIFUSE team will hold an online workshop next summer to share learned methodologies with faculty at other institutions. The workshop is being designed to spread the techniques developed for data science module creation, as well as the modules themselves, positioning the College as a true pioneer in the field.

“It’s really neat to see this effort expanding at Dartmouth,” says Parker.

—Julie Bonette is contributing editor to Dartmouth Engineer