Data Science and the Church of Jesus Christ of Latter-day Saints’ BYU-I.
What does a religious institution and data science have in common? Much more than you think! Our Quantcrunch Report continues to find fascinating examples of just how pervasive data science can be. Brigham Young University – Idaho offers a freshly minted data science degree program driven by passionate, forward thinking faculty who are members of The Church of Jesus Christ of Latter-day Saints. This degree program may be one of the best kept secrets in undergraduate data science. Find out why!
A University Guided by a Church with a Mission
Brigham Young University – Idaho (BYU-I) was founded and is supported by the Church of Jesus Christ of Latter-day Saints (hereafter “The Church”), whose members are often referred to as LDS or Mormon. As part of BYU-I’s stated mission to develop disciples of Christ to be leaders in their homes, Church and communities, the university promises to its students to provide,
“a high-quality education that prepares students of diverse interests and abilities for lifelong learning and employment.”
BYU-I is unique in that it provides only undergraduate level degree programs to over 20,000 students on campus. Its faculty does not conduct research and most of its students are Church members. BYU-I has roots as a community college, so its focus is on addressing the needs of all types of students.
It is in this context that BYU-I’s data science undergraduate degree program has been developed over the last two years.
We recently interviewed one of the founders and the Director of BYU-I’s data science degree program, Professor J. Hathaway. Professor Hathaway, along with other faculty in the Computer Science (CS) and Computer Information Technology (CIT) departments, launched BYU-I’s bachelor degree program in data science in 2017. Professor Hathaway walked us through what it takes to develop an undergraduate data science program from scratch and get it approved by top Church leaders and Church educators. He also gave us his personal views on how best to prepare students graduating from a “university in the middle of nowhere” for a career in data science and analytics.
How the BYU-I Data Science Degree Came to Be
After a ten year career working for the Pacific Northwest National Laboratory, a Department of Energy research lab, Professor Hathaway started teaching in the Math Department at BYU-I in 2015. Back then recalls Hathaway,
“I had been fascinated with how to do an undergraduate data science curriculum.
So, when I got here, my colleague and I proposed it. “
To achieve this however, there was work to be done. The university had some building blocks of course work that could contribute to Professor Hathaway’s vision for the data science curriculum, but nothing obvious.
“We had a business analytics degree that consisted of a bunch of SQL and theoretical economics classes. This degree was managed out of the Computer Information Technology Department. So along with another faculty member, we looked at how we could leverage statistics courses in that program. Those initial discussions evolved into the data science degree.”
Eventually other department Chairs got very involved, which was key to getting this cross disciplinary degree program launched. Of note here is that at other universities, faculty who are passionate about starting an undergraduate data science curriculum typically would sacrifice some level of prestige to create such a program, because they may not be doing as much research and publishing while doing so.
However, explains Professor Hathaway, BYU-I is almost antithetical to faculty doing traditional academic publication research. So, because faculty from different departments were not siloed and focusing on their field-specific research and publishing, they were able to fully collaborate on the development of the data science curriculum. The degree is in fact co-run between the CIT, Computer Science and Math departments.
Explaining how this helped optimize the data science program Professor Hathaway emphasized that BYU-I faculty motivations align with employer and student interests, rather than research and publications.
“I just want to solve business problems. Give me something to do I want to figure out your problem. I don’t need to publish the results for myself. That worked well for our collaboration across departments because we didn’t have to silo ourselves into journals to try to get tenure.”
At BYUI getting a new data science curriculum in place required the additional step of getting approval from the university board which includes the Prophet – the leader of the Church of Jesus Christ of Latter Day Saints.
Professor Hathaway explained that one challenge to this was that the Church leadership is very hesitant to create new classes on campus because they are worried about teachers creating pet classes with only two or three students enrolled.
“I knew we needed a few classes based on my work in industry and the recommendation from the American Statistical Association stating what undergraduate data science should be.”
After spending a semester putting together a ten-page proposal for the university board, Professor Hathaway recalls that he and his colleague joked that they should simply write a summary for the Church education board on the back of a paper napkin along the lines of,
“Listen, there are so many jobs, you do not need to worry about this. These students are going to make money. We are not talking jobs at $30,000 a year, but between $50,00 and $70,000 a year. They’re going to be employable. There’s no talent supply and there are 40,000 jobs a year.”
Joking aside, their proposal worked. The board approved the data science degree program rather quickly. BYU-I now offers bachelors and applied associates degrees, minors, clusters, and certificates in data science. The university expressly created this “nested degree” flexibility to provide a credential in the event that a data science student must leave the university before earning their 4-year degree.
How the Data Science Degree Program Has Developed
Back in 2017 when the data science degree program launched about 20 students registered for the major. In just two years the program has grown to about 125 students and continues to gain momentum with new courses and class sections added every semester.
In the early days the data science major pulled students from other degree programs, but surprisingly not many computer science majors. That said plenty of CS students take a lot of the data science classes, such as the core classes on Data Wrangling, Applied Regression and Machine Learning.
The students that are switching majors to data science are coming from financial economics, sociology, psychology, math and sometimes physics. Professor Hathaway explained why he thinks this is,
“Traditional undergraduate math degrees struggle to prepare students for jobs. Traditional physics degrees have this problem too. A lot of traditional math curriculum is kind of soft around the day-to-day skills needed in industry… So, we’ve pulled a lot of students from our math degree because students started to realize, ‘wait, your data science students get jobs? They get to solve real problems?’”
Math for Data Science
Math skills are of course a huge piece of any data scientist’s job. As a math instructor and statistician, Professor Hathaway has a clear view on math as it applies to the field of data science. Specifically, he looked to the American Statistical Association (AMA) for guidance.
“The AMA said data science students should not go through traditional math classes and that we should create different courses for them. I made a conscious choice, and the computer science co-faculty was OK with this as well, to not use traditional math courses in the data science degree.”
When BYU-I first launched its data science degree, there was in fact no math requirement. This meant that students didn’t have to take calculus or linear algebra. However, they could choose to take these courses if they wanted, and many of them did. But as Professor Hathaway explains, some of the less capable students at that point were missing critical math skills.
The fix? Create a new math class just for data science students.
“Our new data science degree program starts in spring 2020 and we’ll be offering a math class called ‘Applied Calculus for Data Analytics’. Our major will then require calculus and strongly suggest linear algebra.”
Revamping Traditional Courses for Data Science
New data science curricula are often criticized for simply being a repackaging of existing statistics and other courses. This is not the case at BYU-I. BYU-I has had massive growth in existing courses that were completely revamped to be more “data science friendly”.
For example, the university’s traditional course on regression was highly math-focused and had very low enrollment. So, the data science faculty pushed hard asking each other, “Do we have to make it that math based? Can we get students more into applied projects?”
They then shifted the focus of the regression class. As a result, enrollment exploded. The course used to have just ten students enrolled once a year. Now fifty students enroll in it every semester and BYU-I is considering offering two sections of the class each semester.
Similar changes were made to the machine learning class. It used to be offered once a year. Now it’s offered three times a year, and sometimes with two sections a semester since it now has 40-50 students enrolled in each class each semester.
In addition to shifting computer science courses, a few other degree programs have begun leveraging the new data science courses. Accounting is one of them. That department decided that its students need more analytical capabilities and should know some programming. This coming spring, the accounting degree program will go live with a data science emphasis. Accounting majors will be able to take fifteen credit hours of courses using Python, R and SQL and they will touch on different aspects of data science analytics as well.
Other degree programs that have begun leveraging the data science curriculum include the bioinformatics degree and the business analytics degree programs.
New Data Science Courses
Reflecting on the degree program approval process, Professor Hathaway explained that he was pretty sure the university board wasn’t going to approve a new degree with three new classes. So, he and his colleagues prioritized new data science classes.
The first class they created was in his area of specialty, data wrangling and visualization. In that class students learn and use the Tidyverse within R. Professor Hathaway explains the rationale for this,
“That was the first class that was built for this degree to kind of ‘beat’ the students into realizing, OK, you’re a data scientist. You will need to think, communicate, and move data.”
During the time that the data science curriculum has been evolving, as a reflex to industry needs and the influence of Python in industry, the Computer Science department completely rebuilt all of its courses and degrees. That department is launching its new course structure in spring 2020. One guiding principle was to allow the computer information technology, computer science, data science degree students and the entire university to all start learning programming with Python.
The idea is that everyone is going to learn Python first to get them into the programming mindset. The data science curriculum was able to grow further with this push for Python. It will add three new data science courses (Data Intuition and Insight, Data Science Programming, and Big Data Programming and Analytics) starting in the 2020-2021 school year.
In the process of that course overhaul faculty created three new data science courses:
- Data Intuition and Insight – This course will try to avoid programming. It will instead teach students about the art of how data is stored for data science, in the sense of how to organize data in rows and columns to think about doing visualizations. Students will also learn visualization principles and how to use visualization tools.
- Data Science Programming – This course will take students with some background in Python and introduce them to the Python packages (Pandas, NumPy, Seaborn, Sci-Kit Learn) and programming associated with data analytics.
- Big Data Programming – Right now this course is a concept but will likely be launched within the coming year. Professor Hathaway explained the rationale for this course,
“What happens in interdisciplinary degrees like data science a lot is that you put students in all these courses and their like ‘Oh, SQL, R, Python, yeah, I get it’. But then they build an Asteroids game with Python and they don’t realize that Python is a tool for data science and don’t learn how to leverage it. They also don’t grasp that R, Python, and SQL are used together in many projects. We’re developing a class that will start in 2021 to get them into cloud-based technologies and which teaches them how to use SQL, R and Python to solve a business problem at scale.”
As faculty started the data science degree program they reached out a lot to employers and tried to connect to industry. They feel that part of the path to developing the major is trying to stay really connected to industry and “doing what they want done”. But this presents challenges as Hathaway explained,
“Obviously I don’t get to connect to Google because they’re not going to go to some college in Idaho that has undergraduates in data science.”
Another challenge is educating incoming freshman on the data science major. The problem, Professor Hathaway says, is that high school graduates generally have no idea of what data science is or its purpose. They know what scatterplots are and understand that there is a lot of data now, but they don’t realize that there’s an undergraduate degree in that. So typically, freshman don’t come to BYU-I and declare a data science major.
Numerous Opportunities for Data Science Projects and Work Experience
What we found to be distinguishing about BYU-I’s data science degree is the multitude of practical experiences that students have access to during their undergraduate studies. BYU-I provides several avenues for students to pursue practical work on real-world data challenges that organizations are facing.
- Data Science and Statistical Consulting Class
This class has been revamped for data science students and now has twenty students a semester working for three to four large companies each semester, some of which are publicly traded. Many companies give students projects that involves “pretty grungy data work”. Some projects have analytics in them too. Students work on these projects in teams of four or five to solve the business problem. They then present to the companies at the end of the semester. Of this experience Professor Hathaway feels,
“That experience gives students the kind of skill that I think employers want. And I try to stay out of it. I teach them, but you know, I say, ‘Listen you have to solve this problem. In the previous classes I told you what to do, but you need to tell me what you’re going to do now. I’ll tell you if that’s good or not, but you’ve got to work out your plan. Don’t expect me to tell you what to do.’”
He adds that his philosophy for this course is that,
“They’ve got to learn to think through the problem and realize that their boss may not know how to think through the data part of the problem. These companies know that they need help, but they don’t know how to do all this stuff. “
- Senior capstone project
The program requires all graduates to do a senior project. For this, they interact with one faculty advisor for one semester and do a data project on their own which they then must submit through GitHub.
- Data Science Society projects
If students are seeking more opportunities to work on real data sets, they can join the Data Science Society (DSS). When Professor Hathaway started the DSS in 2017 it had about three students and it was “a little pitiful”. But then he worked with student leadership and together they decided that it wasn’t going to be useful for students to sit in on another lecture. So the DSS decided to find projects to work on.
Now DSS members work on smaller projects with companies. They work on about eight real world projects throughout the semester. About sixty students show up to meetings every Wednesday night a 6pm to get into a project team and work on business problems. Professor Hathaway says that the DSS would love to work with your company if you are interested. Simply contact email@example.com to do so.
Students regularly participate in internships and are required to complete one full-time off-campus internship. Students have done internships in many locations in the mountain west, as well as on the west and east coasts. Companies students have interned with include Kraft-Heinz, Honeywell, Moog Aircraft, USAA and InterMountain Health Care.
5. Research and Business Development Center Projects
A unique aspect of BYU-I is that it does not accept external funding from (other than Pell grants), government or grants. To manage the impact of this, the university works with a non-profit that operates independently from the Church called the Research and Business Development Center. The center offers students the opportunity to engage in applied learning projects (ALPS).
This fall semester BYU-I launched its first data science research group in the center. The data science group at the center has already worked on three or four projects. This team interacts with the Data Science Society and consulting class students and will invoice corporate clients. Professor Hathaway is hoping to create several strong relationships with businesses so that students are doing work for a consistent set of companies.
The net result of all of these opportunities is that students can potentially graduate with a portfolio of real world data science projects and 1-2 years of work experience, having spent a semester or more of credit hours doing experiential data science along with the coursework on machine learning, regression, applied statistics and SQL.
“If a student is really committed to their data science degree, they could have a full portfolio on their resume that says ‘I did this for companies X, Y and Z’. These are real world projects and they are not getting credit for them. They’re just showing up to do this work because they are committed.”
In light of this plethora of hands on learning opportunities, Professor Hathaway downplays the perceived need for a Master’s Degree in Data Science. He suggests that some of the aura around needing a master’s degree comes from Data Scientists themselves who are playing undergraduate degrees down at some level. He recalls attending an R Studio conference a couple of years ago and that one of the conference speakers, a Data Scientist, said, “You can’t get an undergraduate data science degree and do anything valuable with it. It’s a waste of time.”
Professor Hathaway says that on some levels he understands this perspective, but says that the reality for his students is different.
“If you’re going to teach data science like a traditional degree with a couple of programming classes then yeah, it is a waste of time. But, we’re going to teach our students that they need to have data empathy and that they need to realize how to communicate and interact, and they need to build the skill sets in their undergraduate program that many are typically developing the first three years of a normal data science job.”
On the Importance of Communication Skills for Data Science
Communication skills are a focal point for BYU-I’s data science program. Professor Hathaway explains that in his experience professionals working in STEM fields sometimes falter when they talk to people because they don’t value communication as part of what they do. At BYU-I some of the science teachers even questioned the need for communications courses in the data science major.
Professor Hathaway insists that it’s very important to the degree.
“I think that’s a beautiful thing about data science too, is to realize that humans are consuming this. This isn’t just a computer interaction. So, we have a communications part of the degree. Students have to take six credits or two communication courses, such as design thinking courses.”
He emphasizes that design thinking was created to try and help engineers have empathy and to realize that problems needed to be solved with groups of people, and “not you in your office going ‘this is the optimal way for me to push my foot on this pedal.’”
In BYU-I’s design thinking courses students get an experience that is not data science at all. They might build a board game for 14-year old’s that will teach them something good or moral. Then they may have to do research to figure out what 14-year-old kids are like now, talk to some of them, figure out what game mechanics are and work in a group to solve “thick data problems” such as interviewing people.
In these classes data science students are also collaborating with students pursuing other degrees. They get the opportunity to see how others approach problems, think differently and want to discuss other things. In this sense they learn to interact and be more engaged with others outside of the field.
Where BYU-I Data Science Graduates Get Jobs
Given all the project work opportunities, it’s not surprising that job hunting is going well for BYU-I data science graduates. Professor Hathaway enthuses,
“We’ve had 12 graduates thus far and they’ve all gotten jobs in this space. A couple of them have gotten pretty high end data science jobs, maybe half have gotten jobs at mid-sized companies who want to get into data and they need somebody that can program better and who is not just an accountant or a business finance major.”
That said, because he is a statistics teacher with a master’s degree Professor Hathaway is starting to get more students asking for advice about graduate school for data science. He explains his hesitation to recommend this route,
“My problem with data science graduate programs is that some may be just like our undergraduate degree. Like undergraduate programs, they are so variable you really can’t tell if certain graduate data science programs are good or not. I often recommend to data science students to go into graduate school for statistics, IT, CS, or economics.”
Job possibilities for undergraduates are becoming more high profile. Recently a Fortune 500 company (alas, not Google) has taken an interest in BYU-I graduates after accepting a student intern there and offering her a job. Says Hathaway,
“They called me recently and said, ‘Tell us more about your undergraduate program and what this program is all about. We think we’d like to interact more with your program.’”
Professor Hathaway says that this new interest is the result of careful thinking about how to craft the curriculum and prepare students.
“We’ve carefully tried to build out the program to fill a gap. A $150,000 Data Scientist hopefully works really fast, hopefully has a high-end analytics ability and knows automatically what tools they are supposed to use and what to do and what to check. My perception of our undergraduate data scientist is that this is a person that can talk to business people reasonably well, can code data and realize ‘I can do this and I can do that to solve a problem, and I can pull data from here. Then I can write a report and visualize it to talk to you about what’s going on here so that we can make the next decision.’”
So what is happening in fact, is that companies are realizing through the BYU-I interns that they can take some of the the work that a $150,000 Data Scientist was traditionally doing and get it done with a $60,000-$70,000 undergraduate. Professor Hathaway further elaborates on what employers are thinking,
“I think there’s a money-making structure to this. Companies have caught on to the fact that we push our students to work with complex data upfront in the curriculum so that they have the confidence to say, ‘OK, I can work on this, I’ve seen real datasets before.’ They see that we push them down that path.”
He notes that another large hospital network in Utah is following the same path. At first, they were a little anxious to take on BYU-I interns, but they were interested in the data science degree. They’ve now taken three interns and hired one of them as they realized that these undergraduate students can actually help solve data problems.
In particular, says Professor Hathaway, BYU-I students are adept at visualization and communication, which draws the attention of the employers they intern with. They’re able to communicate.
Despite these successes there is still work to do educating big companies that come to campus for other reasons about BYU-I’s data science graduates. Professor Hathaway says that it’s hard because the degree is new. He cites the example that BYU-I has one of the biggest automotive degree programs in the nation. Ford and Tesla recruit from it, but they haven’t yet understood that they could also recruit from BYU-I for data science and analytics roles.
The challenge is to break through and make them realize that BYU-I has a solid data science undergraduate program. As Professor Hathaway puts it,
“We have a pipeline of students for data science, this is not a one-off thing. Companies like Walmart and Ernst & Young hire many students every year from BYU-I. We want to communicate that we have a supply of students that have analytics capability and that can do data work for them.”
One challenge students face in their job hunt right now is the sheer number of job applications some have to make to get a job, with students applying to as many as 100 jobs before being hired. Professor Hathaway is trying to support graduates with this challenge, encouraging them to apply to roles that require experience. He advises them, “If it says three years experience needed, you’ve got three years, you’ve been doing all this consulting work. Just apply and describe what you’ve done.”
The real issue he says though, is that the process of hiring undergraduates in data science is “just so nebulous and not really clear”. He has observed that students have to filter through a lot of messy job descriptions that are unclear about what the specific requirements for a job title are. Observes Hathaway,
“Frankly many employers are illiterate when describing what their data science needs are, so they’re not getting what they need for their business problem. Some companies don’t even know that undergraduates can pull off the work.”
That said he does think BYU-I is starting to turn a corner and will see more interest from large employers soon as his department continues to work with the university to create awareness.
Data Science and the Church of Jesus Christ of Latter-day Saints
Besides approving the data science degree, the Church leadership has a significant influence on the curriculum and its outcomes.
Diversity in Data Science
One might not think of Latter Day-saints as being diverse. However with over 16 million members worldwide and over 65,000 missionaries, the Church has a global reach and this is reflected on campus.
The Church is in fact contributing to diversity in the field of data science, as Professor Hathaway explained,
“The funny thing is we are in the middle of nowhere. We’re an hour and a half south of West Yellowstone. Yet our campus is a pretty diverse spot because of how the Church pushes kids to come and get educated.”
To that effect, tuition for Church members from anywhere in the world is pretty affordable at about $3000 a semester. The university also accepts some nonmembers with connections to the Church. Student diversity comes from this education push and from BYU’s unique BYU Pathway program whose mission is to help low income, under educated people get into education.
Apparently, it’s working. Approximately 25% of BYU-I’s student body is non-Caucasian. These diverse students are also not necessarily American born African Americans or Hispanics, for example. They’re coming from other countries like Brazil and South Korea. For example, about 10 students from South Korea are in the data science degree program.
Data Science Job Opportunities
The Church also provides employment opportunities for data science students. With its long history of genealogy databases and analytics (Ancestry.com was founded by two LDS members) it is, according to Professor Hathaway, “very tech savvy”.
“Our Church is excited to take our data science students as well. They’re a big institution so they have traditional IT departments and business analysts. We’re trying to work with them to figure out where our students should go. They are excited to leverage our students for their analytics and foreign language skills.”
The BYU-I campus also hires students to do work. BYU-I typically has about sixty data science majors on campus per semester and 30-40% are working for the university as analysts using their data science skills.
In addition to job assistance, Professor Hathaway believes that Church missions give students a slight leg up (Most young adults in the Church go on missions for 18-24 months). He pointed out that many of his data science students have spent two years “getting kind of beat up learning how to talk to people and be sociable, learning how to maintain a conversation with someone for 20 minutes.” He adds,
“Without having a college degree LDS students have been taught how to maintain eye to eye contact and long conversations.”
Professor Hathaway summed up the Church’s involvement in the data science degree,
“The university and Church leadership is actually excited that we are doing this cross-discipline work and getting students jobs. That’s what’s important to them.”
On the Future of the Data Science Field
Professor Hathaway has been around just long enough to have perspective on the data science hype, which he offers,
“I actually think that it’s like the computer science history back in the 70s and 80s. Back then you couldn’t be a computer scientist without having an undergraduate degree in physics or math. People would say, ‘Oh, to be a computer scientist go get a math degree and then get a Master’s Degree in Computer Science’. But then over a few decades it shifted.”
He described how when he worked at the Pacific Northwest National Laboratory ten years ago you had to have a PhD to work there. The exception to this rule was for statisticians, of which he was one. They could “get away” with a master’s degree. He points out however, that,
“Most of the computer scientists had a bachelor’s degree – and that’s at a research lab. So, we’ve watched that whole space go, ‘um, it’s not really clear you need a graduate degree in computer science to do most of the work. I think that’s going to happen with data science too. Industry and corporations are not quite realizing the extent of this wave of undergraduates that’s about to hit. The wave will be bigger than people think!”
We’d like to thank Professor Hathaway for his efforts and enthusiasm to educate us on his very practical vision for BYU-I’s degree program in data science. If you’d like to learn more about Brigham Young University – Idaho’s data science program check it out here or reach out at firstname.lastname@example.org. For additional background on the unique innovation and philosophy of BYU-I’s data science program check out minutes 9-15 of this video.
If you enjoyed reading this Quantcrunch Report, sign up for our monthly blog digest to receive more data science stories, thought leadership and news from QuantHub.