Roundtable on Data Science Postsecondary Education A Compilation of Meeting Highlights (2020) / Chapter Skim
Currently Skimming:
4 Meeting #3: Data Science Education in the Workplace
4 Meeting #3: Data Science Education in the Workplace
Pages 31-45
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 31... ...
Census Bureau are expected to be unbiased, statistically accurate, delivered quickly at low cost, useful to determine causality, reproducible, transparent, and protected. While striving to meet these expectations, statistical agencies confront many challenges, including greater than expected costs and lower than expected response rates for surveys, complex information requests, competition among data products and questions of product validity, new data sources and methodologies, and policy requirements.
|
From page 32... ...
The Census Bureau is evaluating program use cases to determine which skill sets will be needed by both current and future employees, as well as how projects will be funded. According to Prevost, current knowledge gaps include data science, business/data analytics, reproducibility, software design and engineering, data storage and retrieval models, and operations research.
|
From page 33... ...
The Commerce Data Academy2 is an internal upskilling data science education initiative that relies on the Commerce Data Service, as well as extra-governmental instructors from organizations including General Assembly and Data Society, to train Department of Commerce colleagues in data science, data engineering, and web development skills. After training more than 1,500 Department of Commerce employees in 35 courses 1 The website for the Federal Statistical Research Data Centers is https://www.census.
|
From page 34... ...
Initiated by the Office of Science and Technology Policy during the Obama Administration, Fellows in Innovation3 reaches programs across government, representing 400 fellows and 30 agency divisions. Zachary noted that this program allows data professionals to apply their often underutilized technical skills to a policy problem, as well as to transfer these data science skills to their teammates.
|
From page 35... ...
Offering workplace data science training not only improves employee performance but also may increase employee retention, according to Plachy. IBM created a Data Science Profession to encourage data scientists to continue to train and develop their skills; it uses "open badges" that contain metadata representing "skill tags" and accomplishments, both to signal and verify employees' skills and to improve social connections among colleagues.
|
From page 36... ...
The Consulting Perspective Ashley Lanier and Ashley Campana, Booz Allen Hamilton Lanier and Campana noted that the need to fill knowledge gaps in employee education is not a problem unique to the field of data science. At Booz Allen Hamilton, while employees without data science training need to learn how to use tools efficiently and to analyze and share data, employees with data science specialties need to learn "consulting skills" such as communicating, storytelling, working with clients, working in a team, understanding an audience, and choosing the right approaches.
|
From page 37... ...
These assessments have also revealed that participants are more incentivized by the opportunity to make a difference solving real problems using real data sets than by the opportunity to earn social media "badges" or prize points for their work. Gross suggested that Booz Allen Hamilton publish future assessment results, as doing so could aid the larger data science community in its development of training.
|
From page 38... ...
Lanier noted that Booz Allen Hamilton currently utilizes research collaboration sessions, rapid innovation workshops, and design thinking exercises to facilitate internal problem solving. DATA SCIENCE TRAINING IN THE WORKPLACE: EXECUTIVE EDUCATION Executive Education Online Brian Caffo, Johns Hopkins University Caffo described Johns Hopkins' Data Science Specialization,4 delivered via Coursera, which includes the following courses: The Data Scientist's Toolbox, R Programming, Getting and Cleaning Data, Exploratory Data Analysis, Reproducible Research, Statistical Inference, Regression Models, Practical Machine Learning, Developing Data Products, and a Capstone Project done in collaboration with industry.
|
From page 39... ...
2. Building a Data Science Team -- Overview of differences between types of data scientists and data engineers and how they can work together effectively.
|
From page 40... ...
Executive Education in Business Schools Claudia Perlich, Dstillery and New York University Perlich described a course she offers at New York University titled Data Mining for Business Intelligence that offers two tracks for master's in business administration (MBA) students: the technical and the managerial.
|
From page 41... ...
Census Bureau, expressed concern about offering two separate tracks for the course, since the managerial-track students may not receive the same critical assessment experience as the technical-track students. Perlich responded that it would be difficult to cater to two audiences if faculty delivered this content via a single course, and she added that even if the tracks did not exist, students would likely self-select a course that best meets their knowledge and needs based on the syllabus content.
|
From page 42... ...
Abowd remarked that, when it is possible to hire new employees with different skill sets, the government needs help creating job descriptions that attract appropriate candidates. McKeown encouraged organizations in the public sector to weigh the benefits and drawbacks of hiring new employees versus upskilling current employees and added that recruiting and retaining individuals in government jobs that pay less than industry jobs can be challenging.
|
From page 43... ...
She posited that government agencies may face similar obstacles to data sharing because some data assumed to be open could actually include copyrighted material. Abowd suggested that faculty check the Census Bureau data application program interface7 for data they could freely use within the classroom.
|
From page 44... ...
Patrick Riley, Google, responded that while people working in traditional technical fields talk predominantly with other technical people, data scientists need to be able to explain difficult concepts to nontechnical audiences. Perlich agreed that students have to learn to frame problems clearly for nontechnical audiences.
|
From page 45... ...
Perlich added that while there is no shortage of good will, there is a crucial lack of project management, especially in volunteer programs attracting data scientists. She thinks that a model to ensure that people with the right skill sets are brought together and that volunteers are doing work related to their areas of expertise is needed.
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.