Home » Additional Features

New Graduate Data Science Programs, Continued

1 December 2019 3,598 views One Comment
The proliferation of master’s and doctoral programs in data science and analytics continues, seemingly due to the insatiable demand of employers for data scientists. Amstat News started reaching out several years ago to those in the statistical community who are involved in such programs to find out more. Given their interdisciplinary nature, we identified programs involving faculty with expertise in different disciplines—including statistics, given its foundational role in data science—to jointly reply to our questions.

University of Pittsburgh

MS in Biostatistics with Concentration in Health Data Science (HDS)
Year in which first students graduated/expected to graduate: 2021
Number of students currently enrolled: Fall 2020 will be the inaugural class of new HDS students. Two BIOST MS students transferred to the HDS concentration when it was made available to current students this fall.

Ada O. Youk is associate professor of biostatistics and director of the master’s program in biostatistics.

    Describe the basic elements of your data science/analytics curriculum and how the curriculum was developed.

    This is an in-person program with 40 credit hours. Prerequisite coursework includes two semesters of calculus and exposure to basic programming. Students in our program can be full time or part-time. The curriculum consists of courses in the fundamentals of statistical theory and applications, programming languages (e.g., SQL, R, SAS, Python), data science, machine learning and database management, a statistical consulting practicum, epidemiology, and public health. The program also includes a culminating capstone course to prepare a thesis involving innovative data analysis or related to an internship experience. Opportunities for internships and assistantships are available as hourly employment.

    The program was developed with substantial input from faculty in statistics, computing and information sciences, bioinformatics, and human genetics.

    Graduates of the MS Program in Biostatistics will be able to do the following:

    • Address health problems by appropriate problem definition, study design, data collection, data management, statistical analysis, and interpretation of results
    • Demonstrate mastery of the theory underlying statistical methods
    • Implement and utilize appropriate statistical methods
    • Effectively communicate results of biostatistical analyses to scientific and lay audiences
    • Apply research design principles to problems in public health
    • Recognize strengths and weaknesses of approaches, including alternative designs, data sources, and analytic methods
    • Determine the data best suited to address public health issues, program planning, and program evaluation

    In addition, students in the HDS concentration will be able to do the following:

    • Apply data curation and data management techniques such as data munging, data scraping, sampling, and cleaning to construct informative, usable, and manageable data sets for meaningful analyses
    • Apply methods for big data and machine learning to reveal patterns, trends, and associations including visualization
    • Effectively use a programming language (e.g., R and/or Python) for data management and statistical analysis

    What was your primary motivation(s) for developing a master’s data science/analytics program? What’s been the reaction from students so far?

    We have successfully trained students for many years in traditional biostatistical theory and methods. However, the field of biostatistics is constantly evolving. Advances in technology in biomedical and public health research are generating complex high-dimensional and large-volume data. Standard biostatistical and computational methods are not always feasible, or even appropriate, for analyzing and interpreting such data. To address the needs of modern high-dimensional science, computationally intensive statistical methods have become a fundamental part of scientific research and a critical component in advancing public health. We created our concentration in health data science to keep pace with scientific needs and train graduate students properly for the evolving job market while ensuring a rigorous statistical foundation with traditional methodologies.

    The response from our students has been overwhelmingly positive. Students are excited to master techniques that will set them up for success in the modern job market.

    How do you view the relationship between statistics and data science/analytics?

    Statistics and data science are complementary. It is our philosophy that you cannot have one without the other. For this reason, we chose to create a concentration in health data science rather than a separate degree in data science. A concentration added to our traditional MS degree keeps a core of traditional biostatistical theory and methods while giving a unique focus in data science techniques combined with a central theme of public health.

    What types of jobs are you preparing your graduates for?

    Health data scientist, biostatistician, bioinformatician.

    What advice do you have for students considering a data science/analytics degree?

    Students should have a solid foundation in calculus, an interest in statistics and programming, and a passion for public health.

    Students who graduate from our program will be able to handle all aspects of “big data”—including curation and management—and collaborate effectively with a researcher’s questions through excellent statistical communication skills and translating between scientific questions and hypotheses and statistical concepts and results. They will also be able to perform robust computational and/or statistical methods and effectively communicate analytical results that are statistically and scientifically valid as well as understandable. Thus, students enrolling in our program will receive a comprehensive education—more so than what a separate degree in data science could offer—that will prepare them well to enter the modern job market.

    Describe the employer demand for your graduates/students.

    Indications from job searches on Indeed, ZipRecruiter, and Glassdoor—as well as discussions with local employers—are that demand will be high. In fact, Glassdoor ranks data scientist as the #1 best job in America for 2019 and Forbes states, “IBM predicts demand for data scientists will soar 28 percent by 2020.”

    Do you have any advice for institutions considering the establishment of such a degree?

    Institutions must be willing to put extensive time and effort into the research and creation of the new program. There must be a shared vision with all collaborating parties on what is best for your students and how to train them properly for a competitive job market.

    We chose to start with our strength (biostatistics) and expand from there. We researched other programs around the country and met with and received input from faculty within our university who had complementary expertise. This allowed us to collaborate with departments within and outside of our school and create a strong cross-disciplinary degree.

    In addition to our new HDS concentration, we also created a genomics-focused cross-disciplinary concentration: statistical and computational genomics.

    American University

    Master of Science in Data Science Academic Programs and Center for Data Science
    Year in which first students graduated/expected to graduate: We will have our first student graduate this December, even though the program only officially started this year, because we had students who converted from a certificate program to the full master’s program and had advanced through their coursework extensively.
    Number of students currently enrolled: 48
    Partnering departments (indicate lead, if any): College of Arts and Sciences, School of Public Affairs
    Program format: In-person; 30 credit hours required; capstone course (Data Science Practicum) required for graduation; full-time and part-time options available

    Distinguished Professor Jeff Gill focuses on Bayesian hierarchical models, nonparametric Bayesian models, elicited prior development from expert interviews, and fundamental issues in statistical inference. He is co-director of the data science master’s degree program at American University.

     


    Jane Wall had extensive software industry experience before earning her PhD at Rice University in computational and applied mathematics. She initiated the data science program and teaches R and data science. She is codirector of the data science master’s degree program at American University.

     

    Describe the basic elements of your data science/analytics curriculum and how the curriculum was developed.

    • Required for admission: undergraduate degree in any major with 3.0 and basic statistics course (no GRE or letters of recommendation)
    • 30-hour program—full or part-time, including nine hours in application track and a three-hour capstone

    In 2016, American University developed a graduate certificate in data science as both a stand-alone program and a complement to other graduate degrees. The certificate consists of four courses (12 hours) in R for data science, regression, and statistical machine learning. This certificate was well received by students and became the technical core as we looked to expand the program into a master’s degree. We formed an advisory board for the data science program prior to launching the master’s degree with representatives of both business and government agencies. This advisory board helped form and improve the curriculum, as well as give guidance regarding internships and employment for our students.

    American University’s master of science in data science program prepares students to acquire, process, analyze, and present complex data. Students will master both the theoretical knowledge and practical skills used by data scientists in academia, industry, and government. Core courses such as statistical machine learning, data science, and statistical programming in R train students to clean, process, visualize, and archive modern data sets—including text, imagery, and biometric data—apply machine learning algorithms to real data; and use the mathematical and statistical language of data scientists.

    We strongly believe data science is not just a technical field. Students need to understand the field their data belongs to in order to ask the right questions and deal with strange results or outliers in an effective manner. They must also be able to communicate their results to the nontechnical audience. To accomplish this, each student will choose an applied field in which they will analyze data and explain the associated issues in that field, culminating in their capstone experience in their final semester. Students take nine hours in their application track during their program. There are nine available tracks for students to specialize in: applied public affairs, business analytics, computer science, cybersecurity, environmental science, finance, international economic relations, investigative journalism, and microeconomic analysis.

    Students are not expected to have extensive background knowledge of mathematics, but are required to attend a mathematics “boot camp” prior to the start of the program. This is an important component of the program because we do not restrict the bachelor’s degree: Any accredited bachelor’s degree is acceptable, meaning that we welcome students with backgrounds in the humanities, social sciences, and STEM fields. The program is designed to provide an “easy on-ramp,” meaning prior advanced knowledge of mathematics, statistics, and computer science is not required. The program guides students who are new to data science using principles from a basic starting point all the way through required cutting-edge content like machine learning, mathematical statistics, and advanced data handling. Furthermore, we eschew traditional academic constraints that raise barriers to convenient entry: There is no GRE requirement and admissions are done on a rolling basis. We believe data science training in the 21st century should not be constrained by conventions established for a different era.

    In 2018, AU established its Center for Data Science to focus on the theoretical and practical research aspects of computer technology, software engineering, computer architecture, artificial intelligence, simulation, modeling, and related topics. The center hosts workshops and talks from leading experts in data science from around the world. Our students and faculty can further their knowledge by participating in these events. Part of the mission of the center is to develop graduate education opportunities. In this case, the center worked closely with the department of mathematics and statistics and the department of government to help develop the data science master’s program. This arrangement has worked out well, as the Center for Data Science is able to provide administrative and institutional support for the master’s program outside of a single traditional departmental structure.

    What was your primary motivation(s) for developing a master’s data science/analytics program? What’s been the reaction from students so far?

    Research in the physical sciences and engineering were the engines of human development in the 20th century. We conquered the atom, traveled to the moon, developed digital computation and the internal combustion air-breathing jet engine. We discovered laser and fiber optic telephony, provided widespread electrification, and other important technical milestones. Conversely, the 21st century is shaping up to be dominated by the social and biomedical/health sciences. Data access and data analysis will play an indispensable part in our progress in understanding social, psychological, and physiological characteristics of what it means to be human. Our motivation is to train people who want to engage with the data revolution that is well underway in modern society. We are excited to provide high-intensity training in a geographic location that is at the epicenter of this rapidly changing environment: More data may be produced inside the Capital Beltway than outside of it.

    Students have been enthusiastic about the program so far. They recognize that they are the first cohort, which means we are listening to their input and the program is evolving.

    How do you view the relationship between statistics and data science/analytics?

    It is a well-known truism that data science is an umbrella term that encompasses statistics, computer science, data processing, business/government analytics, and even some subfields of mathematics. Therefore, it is difficult to see any one traditional academic department “owning” a data science program. Elements of traditional statistics graduate education will always be a core component of data science training. We need to teach probability, sampling, estimation, inference, data reduction, testing, decision theory, design, regression, and assessment.

    This foundation of statistics knowledge will not change. But a standard statistical agenda alone does not help create the type of cutting-edge data scientists employers crave. These topics should be complemented with data wrangling; visualization; presentation; and the machine learning topics of classification, clustering, dimensionality reduction, reinforcement learning, and natural language processing. But expertise and ability in the combination of these skills requires computational skills, especially programming.

    Steve Jobs famously said that “everybody in this country should learn to program a computer, because it teaches you how to think.” Naturally, this is even more true for those who need to turn data into information. Therefore, computer science education will always be a key component of training data scientists.

    The future of data science is not going to be strictly in rectangular data files, data dictionaries, and PDF codebooks. We have already seen a revolution in text analysis, biometric measurement, internet traffic, and large-scale social network data, just to name a few. Size is also a challenge for providers and analyzers of data since users will demand enhanced access to extremely large data collections such as genomics, Facebook, global satellite imagery collections, vector GIS, and internet traffic. Data science, even from its infancy, is much more focused on the problems associated with big data than traditional statistical work.

    What types of jobs are you preparing your graduates for?

    We look at this in two ways. First, our core curriculum of seven courses (21 credits) is focused on the key elements that a data scientist needs in any relevant job: data wrangling, statistical inference, statistical modeling, software expertise, machine learning tools, and communication.

    Since we also want to prepare students to do this kind of work in their area of interest, we have nine applied tracks consisting of three courses (nine credits): public affairs, business analytics, computer science, cybersecurity, environmental science/GIS, finance, international economic relations, investigative journalism, and microeconomic analysis.

    We are also working to add more of these applied tracks over the next few years. In our conversations with employers, we learned that the combination of technical skills and experience working with the exact type of data used in their institutions is a sought-after profile.

    What advice do you have for students considering a data science/analytics degree?

    Data science is such a rapidly evolving field that there is an incredibly varied set of programs offered by universities. Students should consider several important features that will differ widely across programs. First, look at the prerequisites. Some programs will expect a relatively high level of mathematics, statistics, and programming experience before starting classes. If students do not meet these levels, it could be challenging to keep up from the start.

    Second, consider what applied tracks or applied foci are offered. Is there an ability to add applied data science in an area you are interested in? This can be crucial for students who have a specific career goal for after the degree. For example, some data analytics programs are offered through business schools (as done here at American University in a completely different program than ours). This often means all the examples will be drawn from business problems. So, for example, a student interested in applying data science principles to biomedical questions will likely be unsatisfied with this component of the program.

    Third, look at the placement outcomes. Is there a university placement office the data science program works with in a targeted fashion? If the program has been around for more than a few years, what is the placement record? Does the program liaise with local employers? Notice this last set of questions will distinguish in-person versus online programs to a great extent.

    Fourth, when are classes offered? This is less of a concern with online programs, but many students pursuing a data science master’s or related degree want to continue working full time and still enroll in a face-to-face program.

    Finally, look at the backgrounds of the core faculty teaching in the program. They are likely to emphasize concepts that have been important in their research and occupational background. Are these areas important to you?

    When we were designing the MS in data science here at AU, we were circumspect about positioning it away from other currently offered degrees so to have a unique new offering. On the technical scale, with regard to mathematical statistics, it sits between the traditional statistics master’s here and the data analytics program in the Kogod School of Business.

    We offer more traditional graduate statistical content than data analytics and more applied hands-on data analysis than statistics. While programming is an important component of data science, we do not believe in requiring courses on software design, system organization, prototyping, or operating system theory, which are not central to data science in daily practice. Having said that, we have a computer science track in the program that allows students interested in going in that direction to take three courses (nine credits) of computer science graduate courses such as database management, object-oriented programming, and computer networks.

    Describe the employer demand for your graduates/students.

    It is well known that jobs in data science are available in every sector: industry, government, and academia. Every sector in the global economy is using data to make better management and policy decisions. Forbes and Glassdoor ranked data scientist as the number one job in America for the last three years, with a median base salary of $120,931. There is no perceivable end to this trend.

    While we are a young program, we have already been approached by many institutions requesting contact with graduates or near-graduates. Part of this, of course, is our geographic location in the center of the data-generation world. The important part is employers can look at our curriculum for the MS in data science online and see the key skills they require. This is by design, since we spent time with these employers during the design of the program working to understand their needs.

    Do you have any advice for institutions considering the establishment of such a degree?

    Our primary advice is to understand your market. The data science market is different in Washington than it is in Duluth. Data science is the quintessential 21st-century skill, but the way graduates are going to use those skills will differ dramatically based on region, economy, and institution.

    Within the university, it is critical to have senior administration buy-in from the start. We reached a critical juncture when it became apparent there would be no department credit-taking and multiple units needed to cooperate or it would not get done. It also helped that departments outside the founding units could propose an applied track and therefore participate in a positive way without having to be part of the 21-credit core.

    Vanderbilt University

    Master of Science in Data Science
    Program contact: Amanda Harding, assistant director of the Data Science Institute and program coordinator
    Year in which first students graduated/expected to graduate: First class was fall 2019; graduation in spring 2021
    Number of students currently enrolled: 31 in our first cohort
    Partnering departments: The MS program is housed in the Data Science Institute (DSI) at Vanderbilt University, which is a multi-disciplinary unit that serves all schools and colleges at Vanderbilt University and Vanderbilt University Medical Center. The departments of biostatistics, biomedical informatics, computer science, physics and astronomy, and political science have all made significant contributions to the DSI and MS program.
    Program leadership: The director of graduate education for the data science master’s program is Jeffrey D. Blume. The Data Science Institute has two codirectors: Andreas Berlind, associate professor of physics and astronomy, and Douglas Schmidt, professor of computer science.
    Program format: The program is a two-year, four-semester full-time program that includes the completion and presentation of a capstone project.

    From left: Jeffrey Blume, Douglas Schmidt, Amanda Harding, and Andreas Berlind. Photo by Joe Howell

      Jeffrey D. Blume founded and directed Vanderbilt’s graduate program in biostatistics before establishing the data science MS program. His research is on second-generation p-values, likelihood methods, large-scale inference, false discovery rates, mediation modeling, clinical trials, and prediction modeling with missing data.

      Douglas Schmidt is the Cornelius Vanderbilt Professor of Computer Science, associate provost for research, and co-chair of the Data Science Institute (DSI). His research focuses on software patterns and optimizations that facilitate development of mission-critical middleware and mobile cloud computing applications.

      Amanda Harding is the assistant director of the DSI and MS program coordinator.

      Andreas Berlind is an associate professor of physics and astronomy and the director of graduate studies in astrophysics. His research focuses on measuring and modeling the clustering of galaxies in the universe using data from telescopes and computationally intensive simulations.

      Describe the basic elements of your data science/analytics curriculum and how the curriculum was developed.

      Data science is an interdisciplinary field focused on extracting knowledge and enabling discovery from complex data. In our view, data science is a fusion of principles from statistics and computer science applied in domain-specific contexts. As such, critical thinking, communication, problem-solving skills, and domain-specific knowledge are emphasized throughout the curriculum.

      We were fortunate to have the opportunity to create brand-new courses for the entire MS program. Because we did not have to repurpose existing classes, our curriculum is modern and highly focused on the essential elements of statistics and computer science that are relevant in the data science space.

      In addition, we take a fully computational approach to teaching statistical principles. Limit theorems, assumptions, and fitting routines are all introduced, illustrated, and proven computationally. Resampling and Monte-Carlo techniques play a key role in the practice of statistics in our program, and students learn to think about and apply statistical rigor computationally. There is also equal emphasis on statistical inference and prediction concepts, as both play a role for successful data science practitioners.

      Students are trained in three core areas (i.e., computation, data analysis, and practice), with an emphasis on imparting practical experience and sharpening soft-skills (e.g., teamwork, communication, leadership). The three core sequences (consisting of four classes each) are the following:

      1. Computation: focuses on programming, data structures, computer systems, and methods
      2. Data Analysis: focuses on data exploration, analysis, model building, prediction, inference, and algorithms
      3. Practice: focuses on teamwork, communication skills, interpretation skills, case-studies, ethical standards, and cultural awareness

      The computation and data analysis sequences are intentionally not labeled as computer science or statistics sequences. Faculty from multiple departments across Vanderbilt teach in both sequences, and concepts are reinforced across sequences. The traditional discipline-specific allegiances are purposely blurred in the classroom, just as they are in practice.

      Prospective students must possess a BA or BS, submit GRE scores (verbal, quantitative, and analytical writing), three letters of recommendation, and a statement of purpose. The minimum requirements for admission are the following:

      • Single variable calculus (one semester)
      • Programming skill (past courses or professional experience)
      • Experience with data analysis (past courses or professional experience)
      • Evidence of computational experience

      What was your primary motivation(s) for developing a master’s data science/analytics program? What’s been the reaction from students so far?

      We proposed a Master of Science program in data science to leverage Vanderbilt’s existing expertise in biostatistics, biomedical informatics, computer science, and other disciplines, as well as blend Vanderbilt’s big data, quantitative, and computational expertise from numerous departments and programs across the university.

      Our program has just enrolled its first cohort this year. We have 31 students with a wide variety of backgrounds (e.g., computer science, statistics, math, journalism, anthropology, environmental science, quantitative psychology). The students are happy, engaged, passionate, and hard working. We have received a lot of positive feedback and valuable constructive criticism that we are using to improve the program.

      How do you view the relationship between statistics and data science/analytics?

      Obviously, they are intertwined. Our view is that a working and computational knowledge of statistical principles in both inference and prediction are essential for successful data scientists and analytic experts. We have eased the traditional theoretical training and added a requirement that students demonstrate a working computational knowledge of, and an ability to apply, these principles in traditional data analysis and machine learning (prediction) contexts. We emphasize reproducible research, R and Python programming skills, and strong understanding of the principle of sampling distributions and resampling techniques. Our program has a stronger computational flavor than most statistics MS programs, and we tried to do this without sacrificing the introduction and analysis of key statistical principles.

      What types of jobs are you preparing your graduates for?

      We are preparing our students for jobs that require strong problem-solving and communication skills. Our students will be well-positioned to lead interdisciplinary teams to solve real-world problems. Jobs in industry, government, academia, and the nonprofit sector are all possible.

      What advice do you have for students considering a data science/analytics degree?

      Our program is designed to be a professional degree; it is not designed to be a lead-in to a PhD program. Students should learn to think like a scientist and then program, simulate, evaluate, and communicate results accordingly. Students should be comfortable programming large and complex computational tasks in several languages, simulating data that can be used to evaluate methods, summarizing computational experiments in plain language, and working online to make analyses and reporting reproducible.

      Describe the employer demand for your graduates/students.

      Students have not graduated yet, but judging from local interest for interns, we expect strong demand.

      Do you have any advice for institutions considering the establishment of such a degree?

      Focus on interdisciplinary projects and find common ground between disciplines. The impact of your work is directly proportional to how accessible it is to those beyond your discipline.

      New York University

      MS in Applied Statistics for Social Science Research (A3SR)
      Concentrations exist in Data Science for Social Impact and Computational Methods
      Year in which the first students graduated/are expected to graduate: 2016
      Number of students currently enrolled: 47
      Partnering departments: Applied statistics, social science, and the humanities
      Program format: In person; variable credit hours (34–44); optional thesis; mix of full-time and part-time students; mix of traditional and nontraditional students; internship required and many assistantships available

      Marc Scott is professor of applied statistics at NYU and codirector of the MS program in applied statistics for social science research (A3SR). Scott also codirects the Center for Practice and Research at the Intersection of Information, Society, and Methodology (PRIISM), which is closely aligned with the A3SR MS program.

       

      Jennifer Hill is professor of applied statistics at NYU and codirector of the MS program in applied statistics for social science research. She also codirects PRIISM.

      Describe the basic elements of your data science/analytics curriculum and how the curriculum was developed.

      The A3SR curriculum was developed after speaking with colleagues at other universities about their MS programs in statistics and data science. It was also informed by many years that both codirectors have spent collaborating with scholars in the social, behavioral, health, education, and policy sciences. Finally, it was and continues to be strongly influenced by feedback from employers, particularly in the government and not-for-profit sectors.

      The feedback we heard was that graduates of MS programs in data science or statistics too often 1) are not readily able to apply what they learned in the classroom with real data for real problems, 2) are not prepared to learn/discover new approaches or create new solutions once in the field, 3) think they can solve an agency’s problems with a model or algorithm, 4) need better training in data cleaning and reproducible coding practices, and 5) are not skilled at effectively communicating with their employers (this includes listening, writing, speaking informally, and presenting formally). Our course requirements augment a more traditional statistics curriculum to directly address these concerns. In addition, we offer additional courses in modern machine learning algorithms that are increasingly popular in both industry and other sectors.

      All MS students are required to take the following core courses: intermediate quantitative methods; multivariate analysis; data science for social impact; causal inference; a practicum in statistical computing and simulation; multilevel models: nested data or growth curves; survey research methods or experimental design; generalized linear models and applied probability. Additionally, each student selects one of the following concentrations for three to four additional specialized courses: data science for social impact; computational methods; or general applied statistics. Students are also encouraged to explore interest areas with three to four electives in statistics and beyond. Finally, all students complete the culminating experience of statistical consulting while completing a related internship or research project.

      What was your primary motivation(s) for developing a master’s data science/analytics program? What’s been the reaction from students so far?

      Given our success in training doctoral-level social and behavioral scientists in statistical methods, our deans encouraged us to develop an MS program. We decided we would only create a program if we could design it in a way that would be tailored to students and goals not addressed in many traditional statistics programs. We also noted that some of the most popular MS programs in data science downplay statistics training, which leaves their students at a disadvantage when employers assume they understand the basics of inference, hypothesis testing, and uncertainty quantification, let alone critical skills like research design, sampling, and measurement. Finally, given the increasing importance of collaboration and interdisciplinarity, we wanted to focus on skills that directly aid in working with people across fields and backgrounds.

      A3SR has small cohort sizes (20–25 per year) so we have the opportunity to get to know all our students. Our students are smart, motivated, and love to learn new things. They are the types of students who want to probe deeper and who are adept at figuring out how to teach themselves new ideas, methods, and programming languages so their learning extends far beyond the classroom. In addition, our students are supportive and collaborative—often working together to solve problems or complete projects. In fact, during a recruiting session for the program two years ago, one of our students described A3SR as having a “warm, supportive, and loving environment.” Music to our ears!

      How do you view the relationship between statistics and data science/analytics?

      We define data science as the use of data to understand and have an impact on the world. This includes how data are collected (e.g., sampling, scraping, research design, survey design, measurement) and turned into meaningful inputs (e.g., cleaning, measurement, covariate selection, feature creation), understanding the goal of the inference (e.g., causal versus descriptive or predictive), models and algorithms (e.g., flavors of parametric versus nonparametric, supervised versus unsupervised, Bayesian versus frequentist), and presentation of findings (e.g., visualization, written and oral translation for different audiences). Ethical considerations and the soft skills required to collaborate effectively with practitioners and other researchers pervade all these aspects of the data pipeline. We view statistics and computer science as core to the discipline, and contemporary data scientists are encouraged to hone their skills in both fields with the overarching goal of understanding the world and communicating that understanding to stakeholders honestly and transparently.

      What types of jobs are you preparing your graduates for?

      Our students typically work as data scientists, research scientists, or consultants in the public or private sector to improve society at organizations like MDRC, Vera Institute of Justice, New York State Attorney General’s Office, Global Ties for Children at NYU, Center for Education Policy Research at Harvard University, International Rescue Committee, and UNICEF. Additionally, our alumni are doctoral candidates in the following programs: biostatistics at Harvard University; statistics at the University of California, Berkeley; statistics at Hong Kong University; sociology at New York University; and international education at New York University.

      Describe the employer demand for your graduates/students.

      Our graduates have been highly successful at getting jobs after graduating. As our program matures and more organizations have interactions with our students through internships, summer jobs, or project-based consulting, we are steadily increasing the number of employers who actively seek out our students to hire.

      Do you have any advice for institutions considering the establishment of such a degree?

      The biggest challenge for an interdisciplinary program in this domain is balancing building theoretical foundations with developing practical skills. We emphasize both in our courses, augmenting more traditional coursework with a stronger emphasis on programming, learning through simulation, and project-based work in which students use real (messy) data to address scientific research questions.

      Our students also develop skills through consulting projects with NYU faculty and internships. These activities allow our students to gain important experience working on collaborations with closely aligned programs such as those in public health; education; and social, behavioral, and policy science.

      1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
      Loading...

      One Comment »

      • Steve Pierson said:

        See the other Q&A Amstat News articles over the last couple years. They are listed and linked here.