Roundtable on Data Science Postsecondary Education A Compilation of Meeting Highlights (2020) / Chapter Skim
Currently Skimming:
10 Meeting #9: Motivating Data Science Education Through Social Good
10 Meeting #9: Motivating Data Science Education Through Social Good
Pages 122-137
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 122... ...
Welcoming roundtable participants, Kathleen McKeown, Columbia University, commented that many students in data science, computer science, and statistics courses are eager to "give back" to their communities through the practice of data science for social good. She highlighted ethical concerns raised during previous meetings of the roundtable, such as potential bias in machine learning and fair artificial intelligence (AI)
|
From page 123... ...
He suggested that ethics and security be integrated throughout data science curricula and that future data scientists receive increased liberal arts training. He championed the role of 2-year institutions in offering introductions to data science for social good, and he advocated for Congress to support free education from 2-year institutions for all Americans.
|
From page 124... ...
program2 as an immersive fellowship in which aspiring data scientists learn how to map data methods and tools to social problems in partnership with a government agency or nonprofit organization. Gee said that DSSG builds a community of open, ethical, collaborative data science practice through research and development, lectures, workshops, and events.
|
From page 125... ...
First, he explained that because data science has gained popularity, economic power, and cultural cachet quickly, data scientists are often unaware of the potential consequences of their work. Data science education is currently failing in that it is taught at a distance, with clean data sets separated from social context.
|
From page 126... ...
Educators could also provide guidance to employers for incorporating ethics case studies into hiring, apprenticeship, and mentorship opportunities. Taking these steps to improve data science training, Gee said, could render data science as more of a "healing profession with deep purpose and moral authority." Michael Pearson, Mathematical Association of America, asked Gee whether DSSG includes discussions of how data science will inform policy or hold policy makers accountable for data misuse.
|
From page 127... ...
Bhargava said that his classroom is a "playground" where students "flex their data muscles" in a safe learning space. So that other educators can access hands-on data-storytelling activities, this open source content is available through the Data Culture Project.6 Alfred Hero, University of Michigan, wondered how Bhargava achieves a convergence between his course and more traditional data science methodology courses because many students enrolled in the latter may not enroll in the former.
|
From page 128... ...
He addresses similar student concerns through team design, pairing students with different perspectives, learning goals, and work habits. Gee said that some fellows consider leaving the program each year because they dislike the amount of time spent talking with project partners or navigating team politics; however, most ultimately realize that this "messiness" is the benefit of doing clinical practice.
|
From page 129... ...
DATA, DESIGN, AND ENGAGEMENT: LESSONS FROM 30+ DATA SCIENCE FOR SOCIAL GOOD PROJECTS Peter Bull, DrivenData DrivenData7 has worked on more than 50 projects with nonprofits, social enterprises, and corporate social responsibility groups, Bull explained, and it tries to figure out how to solve organizations' problems with machine learning or data science tools, using the data assets that they already have. An organization's problem is posted online, and a community of data scientists proposes algorithms to solve it.
|
From page 130... ...
He described DrivenData's other open source projects, including ookiecutter Data C Science,8 a standardized project structure for doing data science work, and Deon,9 an ethics-checklist generator for projects. DrivenData also engages directly with organizations to solve data science problems.
|
From page 131... ...
TEACHING PEOPLE TO THINK WITH DATA James Hodson, AI for Good Foundation Hodson explained that the AI for Good Foundation10 was established in 2014 after a series of workshops at Stanford University about the s tatus of AI and future innovation. Participants discussed core problems, breakthrough methodologies, and social impacts.
|
From page 132... ...
Hodson said that society should embrace data-driven science; data literacy across campus; cross-disciplinary research and teaching resources; open infrastructure, data, and methods; data innovation hubs; data science for social good; and diversity. The main barriers to achieving these goals are that the methods are often taught independently from the research process; students are seldom taught how to evaluate, clean, and merge data; and the teaching of applied data science in a laboratory setting is too short, too stylized, and has no impact.
|
From page 133... ...
After annotators received training, they began to code the data and to develop a baseline interpretation. Annotators then created descriptions informed by the context of the social media post, and machine learning was used to label data as "loss," "aggression," or "other." Domain experts would then review the labels and help reconcile the interpretations by providing additional context.
|
From page 134... ...
OPEN DISCUSSION Data Science Education Ullman asked the speakers to outline the technical differences between "data science for social good" and "data science." Bull responded that it is important to give all data scientists a concrete process to ask the right questions in order to understand the domain they are working in, especially when it comes time to hand off a solution to an organization. Hodson said that there is great opportunity to make social impact through data science, but that is not the most important part of rethinking data science education.
|
From page 135... ...
. He noted that the foundation tries to unite researchers, students, and community stakeholder groups; it helps external organizations understand where they need advice and interaction and helps researchers understand how theoretical research can be applied.
|
From page 136... ...
When people are forced to work in diverse teams (e.g., data scientists and domain experts) , people step outside of their comfort zones and explore broader issues.
|
From page 137... ...
Treisman noted that the data science community can influence the infrastructures that currently stipulate ethical behavior.
|
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.