UC Berkeley School of Information

You can buy the best hardware in the world, and hire the best mathematicians. You can write brilliant machine learning algorithms. However: if you do not have a way to produce information that is relevant to your organization and successfully communicate it to them, your entire data science department is the functional equivalent of a paperweight that costs more than raw plutonium. So let’s take a minute to talk about organizational structure, information flows, hiring, training, and data’s social signal-to-noise-ratio. Kimberly Stedman Data Scientist Motiga Kimberly Stedman does big data in the games industry. She was originally a field anthropologist, and has lived in five developing countries. Kim has a Master’s in Social and Organizational Systems Analysis. Kim specializes in the design and management of the social systems that surround data technologies. In other words: Awesome! We’ve got a better algorithm running on faster hardware! … Now what? Kim gave a 5-minute Ignite talk on this topic: How to Build an Effective Data Science Department. Kim also blogs as K2. She wrote Brosie the Riveter, a comedic article on gender issues in gaming and STEM.

Protecting privacy and civil liberties cannot and should not be left to the lawyers. Given the rapid pace of technological innovation and the glacially slow development of corresponding legal doctrine, it falls to engineers and technologists to consider privacy and civil liberties issues as they design, build, and sell their ideas. In the course of this audience discussion, we will explore a hypothetical situation involving both business and design choices and consider the challenges in charting the most ethical course. John Grant Civil Liberties Engineer Palantir Technologies John joined Palantir Technologies in September 2010 as the company’s first Civil Liberties Engineer. Previously, John served for nearly a decade as an advisor in the United States Senate. He began his career in the Senate as an aide to Senator Peter Fitzgerald before joining the staff of former presidential candidate and member of the Senate Republican leadership, Senator Lamar Alexander. While working for Senator Alexander on issues ranging from the federal budget to homeland security, John attended law school at Georgetown University. He earned his law degree shortly after joining the staff of the Senate Homeland Security and Governmental Affairs Committee. As Counsel to Ranking Member Senator Susan Collins, John handled the Committee’s intelligence and privacy and civil liberties portfolios. He conducted oversight of numerous programs within the Department of Homeland Security and the U.S. intelligence community as well as investigations into intelligence failures that led to the attacks at Fort Hood and the failed 2009 Christmas Day bombings. As a Civil Liberties Engineer at Palantir, John has worked with customers all over the world, helping them to develop data protection practices that make the best use of Palantir’s privacy and civil liberties protective capabilities.

Making decisions based on correlation can be risky. At Airbnb, a two-sided marketplace with both an online and offline travel experience, engineers have found that experiments provide powerful insights to act upon, by making it possible to distinguish correlation from causation. However, it’s important to be especially thoughtful about the design and implementation of experiments, given the complexity of the Airbnb system. Elena Grewal, data scientist at Airbnb, will discuss how the Airbnb data science team uses experimentation to inform business and product decisions. Learn how Airbnb uses experiments in non-traditional settings, how experiments enable sound decision making across the company, and how the team has learned the potential pitfalls of experiment analysis along the way. Elena Grewal Data Science Manager Airbnb Elena Grewal is a data science manager at Airbnb. She leads a team of data scientists responsible for the online experience and offline travel experience. Her team collaborates with the product team, using experimentation and descriptive analysis to design and optimize the site, in addition to validating future product plans and inspiring those plans with exploratory analysis. The team is also responsible for data-driven products, such as pricing suggestions and search ranking algorithms. Prior to working at Airbnb, Elena completed a doctorate in education at Stanford where she built predictive models of friendships in schools.

The movement known as “open data” started with a handful of governments releasing their data to the public. On one hand, open data has spurred innovative uses of government data, particularly with transit data and apps. But to expand the use and value of open data, we must go beyond mobile apps to leveraging open data as a means to inform public dialogue and decision-making in cities. Learn how San Francisco is migrating from simply pushing data out the door to enabling use of our (that is, your) data. Joy Bonaguro Chief Data Officer City and County of San Francisco Joy Bonaguro is the first Chief Data Officer for the City and County of San Francisco, where she manages the City’s open data program. Joy has spent more than a decade working at the nexus of public policy, data, and technology. She’s worked from the birth of the open data and open government field, spending seven years designing and managing the development of information systems to support planning and decision-making at the Greater New Orleans Community Data. Prior to joining the City, Joy worked at Lawrence Berkeley National Laboratory to help develop technology, cyber and privacy policy working closely with both the National Lab CIO Council and the Department of Energy Information Management Advisory Group. Joy earned her Masters from UC Berkeley’s Goldman School of Public Policy, where she focused on IT policy.

The data science toolkit encompasses powerful approaches for detecting and clarifying patterns in social or behavioral data. But when it comes to the interpretation of those patterns, it sometimes falls short — the data may convey “what” and “how much” with great precision, but it is often silent on “why” and “how.” Complementary research methods can fill in these gaps and paint a fuller picture of the phenomena at hand. At Facebook, we combine data science with qualitative and quantitative research, often iteratively, to gain a deeper understanding not just of what people are doing on Facebook, but why and how. Judd Antin UX Research Manager Facebook Judd Antin is a UX Research Manager at Facebook, where he focuses on bring the theories and practices of social psychology, social computing, and HCI to bear to improve Facebook’s products. Judd and his team employ methods that range from ethnographic fieldwork to big data analysis to understand products like News Feed, Ads, Photos, and Groups. In 2011, Judd was named one of MIT Technology Reviews Top Innovators Under 35 (TR35). Judd holds a Ph.D. from the UC Berkeley School of Information, where his research focused on collective action, social dilemmas, and incentive systems in online collaboration. Judd also holds an MA in Applied Anthropology from the University of Maryland College Park, and a BA in Anthropology from Johns Hopkins University. Andrew T. Fiore Growth Research Team Lead Facebook Andrew T. Fiore leads Facebook’s Growth Research Team, including a program of research to understand and measure the benefits, risks, and barriers to adoption of information and communication technologies in emerging markets. Previously, as a member of the data science team at Facebook, he studied social dynamics in online groups. His past research at UC Berkeley, the MIT Media Lab, and Microsoft Research focused on the design and analysis of behavior in computer-mediated communication systems, including usenet newsgroups and online dating sites. He holds a Ph.D. from the UC Berkeley School of Information, as well as an M.A. in statistics from UC Berkeley, an M.S. in media arts and sciences from MIT, and a B.A. from Cornell University.

A rise in chronic conditions has put a strain on our healthcare system. Treatment for chronic conditions spans time, agencies, and providers. Information systems such as electronic health records should be helping with the challenge of coordination, but too often they do not. My research aims to alleviate this problem by designing health information systems that fit social practices and workflow. In this talk I will describe my research agenda around collaborative reflection – an informal, unpredictable, and adaptive type of decision-making. I have studied collaborative reflection in behavioral and mental health services for children, which are coordinated across clinical, home, and special education settings. Through participatory design I developed Lilypad, a tablet-based information system for collaborative reflection. I then examined the social impact of Lilypad using deployment studies. I will discuss what the Lilypad project tells us about the way health information systems should be designed and integrated within health service organizations, if they are to have a positive impact on stakeholders involved with managing chronic conditions. Bio: Gabriela Marcu is an assistant professor in the College of Computing and Informatics at Drexel University. She obtained her Ph.D. in human-computer interaction this year from Carnegie Mellon University, and her B.S. in informatics from UC Irvine in 2009. Her interdisciplinary research addresses problems in the coordination of health services. She combines computer science with anthropology and design, to develop and study sociotechnical solutions to real-world problems. She has been named a Siebel Scholar, NSF Graduate Research Fellow, Microsoft Research Graduate Women Scholar, and a Google Anita Borg Scholar.

In his doctoral dissertation, Kai Huotari studied how TV live-tweeting influenced the TV viewing experience. He interviewed 45 live-tweeters and analyzed more than 4,000 TV live-tweets sent in the U.S. in 2011–12. The study identified four distinct groups of users live-tweeting about TV programs (fanatic TV live-tweeters, systematic TV live-tweeters, sporadic TV live-tweeters, and active Twitter users), four main categories of TV live-tweets (courtesy tweets, outlet tweets, selection tweets, and analysis tweets), described several TV live-tweeting practices from preparation practices to reading and writing live-tweets and to the use of Twitter functions, and revealed that a TV live-tweeter is an empowered TV viewer who can, by experientializing live-tweeting into his or her TV viewing, personalize and control his or her TV-viewing experience better than before, express him- or herself more fully, and reach a large enough audience and acceptance for his or her ideas.

Jake Peterson discusses the Facebook analytics team and how they perform large scale data analysis, identify actionable insights, suggest recommendations, and influence the direction of the business. The Facebook analytics team serves as the voice of data that drives success throughout the company, including product development, user engagement, growth, revenue, and operations. Learn about their typical day-to-day responsibilities, challenges, and how best to succeed as a data scientist in analytics. Jake Peterson is a data scientist and analytics engineering manager at Facebook and has been working in data science for more than ten years — longer than “data science” has been a term. At Facebook, Jake has led data science for four different Facebook product teams, most recently for the Graph Search product. Prior to Facebook, Jake led analytics functions at several tech startups and spent six years in the direct marketing industry as an analytics consultant at Acxiom. He holds a B.S. in computer science and a B.A. in philosophy from Santa Clara University.

What does the recent battle for patent reform mean for startups and for the future of tech policy? Julie Samuels is executive director and president of the board of Engine, a young and influential advocacy group working to ensure startups have a voice in D.C. Through policy analysis, economic research, and close relationships with policymakers and startups, Engine is helping to elevate the interests of technology entrepreneurship in American policy. Julie gives an overview of the recent battle in Washington for patent reform and talk about lessons learned. She discusses what the battle means for technology and startup policy going forward — and explains why startups and businesses need to stay involved in the fight for change in D.C.

A panel of women data scientists discuss their career trajectories in the emerging and rapidly evolving field of data science. Panelists: Katharine Matsumoto, Data Scientist in Product Intelligence, Salesforce.com Vesela Gateva, Sr. Data Scientist, Eventbrite Emi Nomura, Data Scientist, Jawbone Elena Grewal, Data Scientist, Airbnb Pinar Donmez, Chief Data Scientist, Kabbage, Inc Anno Saxenian, Dean, School of Information (moderator)

Meet Eric Liu and Paul Duan of Bayes Impact, a non-profit organization deploying data science teams to work with civic and nonprofit organizations to solve big social impact challenges. They’ll talk about how Bayes Impact’s full-time fellowship programs bring together domain experts and data scientists from top technology companies and academic institutions and how I School students can get involved.

MIMS student speaker Deb Linton, from the UC Berkeley School of Information 2014 Commencement (May 17, 2014).

From the UC Berkeley School of Information 2014 Commencement (May 17, 2014). Keynote Speaker Nicole Wong is the deputy US chief technology officer, advising on Internet policy and privacy. Prior to joining the Obama administration, Nicole was the legal director at Twitter and vice president and deputy general counsel at Google, primarily responsible for the company’s product and regulatory matters. She is also a former partner at the law firm of Perkins Coie. Nicole is a frequent speaker and author on issues related to law and technology, including multiple appearances before the US Congress regarding Internet policy, censorship and privacy. She has taught media and Internet law and policy courses at UC Berkeley, Stanford University, and the University of San Francisco. She is a member of the advisory board to the Graduate School of Journalism at UC Berkeley and has served on the governing committee of the ABA Communication Law Forum and the board of directors of the First Amendment Coalition. Nicole received her law degree and a master’s degree in journalism from UC Berkeley and her Bachelor of Arts degree from Georgetown University.

When Time Magazine named “YOU” as their 2006 Person of the Year, it highlighted what has been deemed the democratization of the media. The term “Web 2.0” was coined to describe this transformation on the internet, where individual volunteers, not institutions, control its content. But many people share doubts about the hype around Web 2.0 and have different ideas about what's significant, what's trivial, and what's irrelevant. Protagonists, such as Andrew Keen, believe that it is not only significant, but is significant enough to threaten “our economy, our culture, and our values.” Please join UC Berkeley Adjunct Professor Paul Duguid and Andrew Keen in a debate about whether Web 2.0 is truly a threat to our culture. Adjunct Professor Geoffrey Nunberg will moderate the debate. Co-sponsored by the UC Berkeley School of Information, the Berkeley Center for New Media, Mass Communications at UC Berkeley, and the UC Berkeley Library

Experience designers and researchers are working on their most ambitious challenge yet that represents a new frontier in user interfaces: creating a constellation of systems, machines, and people — including wearable gadgets, tablets, smart phones, and appliances — that can communicate with one another in an autonomous fashion. We are constantly adding new functions to gadgets and new devices to the ecosystem without much thought as to their totality or cumulative complexity. In the coming era of ubiquitous sensors and miniaturized mobile computing, designers need to think about how to weave the digital world into our lives at work so seamlessly that we don’t even notice. Already we’re seeing a groundswell of new technologies that insinuate themselves seamlessly into users’ personal lives like the voice and gesture-controlled Xbox, but this is just the beginning. At GE, we want to apply the same embedded intelligence to the world of big iron and people at work to create disruptive experiences, not just products or interfaces, by connecting people with people and people with machines and data. In this talk you will learn how GE Global Research is pursuing new opportunities in analytics visualization, the future of field engineering, wearables, robotics, semiautonomous vehicles, and agents that can be applied to the industrial Internet. Bio: Arnold Lund, Ph.D., CUXP, is the connected experience technology leader and human-systems interaction lab manager at GE Global Research. He was previously the principal user experience lead in Microsoft’s Server and Management Studios and the principal director of user experience in Microsoft’s IT organization. Prior to that, he worked at AT&T Bell Laboratories, Ameritech, US West Advanced Technologies, and Sapient. His research has been recognized by the Society of Motion Picture and Television Engineers (SMPTE), and his work on metrics for assessing user experiences and predicting successful products continues to be influential. Lund is a fellow of the Human Factors and Ergonomics Society (HFES), and served on the HFES Executive Council. He has been elected to the prestigious ACM SIGCHI Academy and recently received the SIGCHI Lifetime Service Award. He has long been engaged in human computer interaction (HCI) and human factors standards, and in the area of accessibility and emerging technology. He chaired the HFES Institute and oversaw the HFES-200 standard and its approval as an ANSI standard. He is a certified user experience professional and served as president of the board of directors for the Board of Certification in Professional Ergonomics (BCPE). Lund has published widely in R&D management and on research in natural user interfaces, and has more than 20 patents. He has taught user-centered design at Northwestern University and the University of Washington.