Data Science newsletter – August 22, 2018

Newsletter features journalism, research papers, events, tools/software, and jobs for August 22, 2018


Data Science News

How Higher Education Can Use Data Analytics to Address Declining Enrollments

insideBIGDATA, Eric Spear


In this special guest feature, Eric Spear, Founder and President of Precision Campus, discusses how colleges and universities are trying to come up with a strategy to combat declining enrollments. Precision Campus is a data analytics software program designed exclusively for higher education. Eric has more than 20 years of data warehouse development experience and has performed all related duties including database management, user requirements gathering, coding, training and data administration.

Seeking a new element in artificial intelligence: trust

EurekAlert! Science News, NYU Tandon School of Engineering


For decades, the cybersecurity community has devised protections to fend off malicious software attacks and identify and fix flaws that can disrupt the computing programs that are central to all aspects of life. Now, a team of researchers from New York University Tandon School of Engineering and Columbia University has received a grant from the National Science Foundation (NSF) to develop some of the first tools to bring those same protections to artificial intelligence (AI) systems.

“There are ways to test and debug computer software before you deploy it and methods of verifying that your software works as you expect it to,” said Siddharth Garg, an assistant professor of electrical and computer engineering at NYU Tandon. “There’s nothing analogous for AI systems, and we’re developing a tool suite that will lead to safer, more secure deployment of the systems used in autonomous driving, medical imaging, and other applications,” he said.

Harris Public Policy announces first-of-its-kind initiative to inform cyber policy

University of Chicago, UChicago News


The University of Chicago Harris School of Public Policy today announced the creation of the new Cyber Policy Initiative (CPI), a first-of-its-kind academic initiative which will advance the emerging field of cyber policy and examine the intersection of national security, politics and technology. The CPI will provide much-needed policy guidance to governments across the globe and the broader policy community concerned with cyber issues.

CPI’s first collaboration is a partnership with DEF CON, the world’s largest and best-known hacker conference, to provide free cybersecurity training to state and local elections administrators from around the country in an effort to safeguard U.S. elections infrastructure.

How Network Math Can Help You Make Friends

Quanta Magazine, Patrick Honner


Studying the structure of existing friendships in your community can help you forge the best connections when forming a new circle of friends.

DSI launches Online Masters program in Data Science

University of Virginia, Data Science Institute


The University of Virginia Data Science Institute is launching an online Master of Science in Data Science, with the inaugural cohort set to start in summer 2019. Through a collaboration with Noodle Partners, the DSI will offer a fully online version of the institute’s sought-after Master of Science in Data Science program.

States move quickly to tap into money for election security

Associated Press, Christina A. Cassidy


Racing to shore up their election systems before November, states are using millions of dollars from the federal government to tighten cybersecurity, safeguard their voter registration rolls and improve communication between county and state election officers.

The U.S. Election Assistance Commission released a report Tuesday showing how states plan to spend $380 million allocated by Congress last spring to strengthen voting systems amid ongoing threats from Russia and others.

Testing Recommended Posts in Feed

Instagram, Info Center


Today, we’re testing recommended posts in Feed, a new way to see content you may like on Instagram. The recommendations are based on the people you follow and photos and videos you like.

You’ll see recommended posts at the end of your Feed, once you’ve seen everything new from people you follow. When you see the “Recommended For You” message, you’ll have the option to view your past posts or keep scrolling for your recommendations. And, if you’re interested in seeing more from the accounts shown in your recommended posts, just tap on the blue button to follow them.

How Nordstrom tapped into a century’s worth of customer data

VentureBeat, Tess Townsend


Nordstrom marketing executive Brian Hovis knows his company is not a tech company. But in an effort to better reach customers and meet their needs, the retailer is getting savvy with artificial intelligence.

The company’s strategy, Hovis explained at the Transform conference in Mill Valley, is to be discerning and agile about when to develop its own internal resources and when to bring in third-party tools.

“We have invested in a data scientist capability in-house. That doesn’t happen overnight. You have to source that talent, onboard that talent, and build that capability,” said Hovis, vice president of marketing for Nordstrom’s full price business.

Google created a fake pizza brand to test out creative strategies for YouTube ads

TechCrunch, Anthony Ha


Google’s Unskippable Labs team has been testing ad effectiveness in a compelling new way: It created a fake pizza brand called Doctor Fork, used stock footage to create 33 ads and then served them up on YouTube and reached 20 million impressions.

Ben Jones, creative director at Unskippable Labs, explained that there are certain axioms among advertisers that are never really put to the test. For example: “You can’t show somebody chewing food and looking at the camera.”

If everyone sees this as an obvious “third rail,” then no one puts it in their ads, so there’s no testing to see if chewing and looking at the camera is really a problem. By creating a fake pizza brand, Jones said his team suddenly had “a very different kind of freedom to be wrong.”

Facebook Identifies New Influence Operations Spanning Globe

The New York Times, Sheera Frenkel and Nicholas Fandos


Facebook said on Tuesday that it had identified multiple new influence campaigns that were aimed at misleading people around the world, with the company finding and removing 652 fake accounts, pages and groups that were trying to sow misinformation.

The activity originated in Iran and Russia, Facebook said. Unlike past influence operations on the social network, which largely targeted Americans, the fake accounts, pages and groups were this time also aimed at people in Latin America, Britain and the Middle East, the company said.

Some of the activity was still focused on Americans, but the campaigns were not specifically intended to disrupt the midterm elections in the United States, said FireEye, a cybersecurity firm that worked with Facebook on investigating the fake pages and accounts. The operations “extend well beyond U.S. audiences and U.S. politics,” FireEye said in a preliminary report.

Swift’s telescope reveals birth, deaths and collisions of stars through 1 million snapshots in UV

The Conversation, Michael Siegel


Launched on November 20, 2004, and orbiting an altitude of 340 miles, NASA’s Neil Gehrels Swift Observatory has three telescopes that monitor the universe using wavelengths of light that are blocked by Earth’s atmosphere. These included the X-Ray Telescope, the gamma-ray-sensitive Burst-Alert Telescope and the Ultraviolet Optical Telescope (UVOT). The UVOT recently delivered its 1 millionth image – data that astrophysicists like me use to gain insights into everything from the origins of the universe to the chemical composition of nearby comets.

Amazon’s latest health care hire brought technology into the clinic – STAT

STAT, Casey Ross


Dr. Maulik Majmudar has spent years toiling on a task Amazon must master to disrupt the nation’s health care industry: getting physicians to incorporate novel technologies into their practices.

This week, he announced he is taking a new job with the ecommerce giant following several years incubating new technologies at Massachusetts General Hospital.

In an interview with STAT Monday, Majmudar, the former associate director of the health care transformation lab at Mass. General, said the job at Amazon offers a chance to drive the uptake of technology solutions that could impact patients worldwide.

Why does AI stink at certain video games? Researchers made one play Ms. Pac-Man to find out

Science, Matthew Hutson


Artificial intelligence (AI) can kick butt in games such as Pong and Space Invaders, but it comes off like a common n00b when playing Ms. Pac-Man (pictured). Now, by making AI play six classic arcade games, researchers are closer to figuring out why thinking machines excel at some games and stink at others, they reported last month at the International Conference on Machine Learning here.

The team developed a new system for visualizing how Atari-playing AIs operate. They chose Atari because the games are relatively simple and a frequent focus for researchers developing “reinforcement learning” algorithms, AIs that learn behaviors through trial and error. An AI “sees” the screen (as an input of ones and zeroes) and at first randomly responds with commands for “left,” “right,” “fire,” and so on, slowly shaping its strategy as it receives points for certain actions. In Space Invaders, the AI moves a ship back and forth across the bottom of a screen while shooting descending aliens and dodging their projectiles.

After thousands of practice games, an AI can best human performance at Space Invaders.

A toolkit for data transparency takes shape

Nature, Jeffrey M. Perkel


A simple software toolset can help to ease the pain of reproducing computational analyses.

Billion-dollar telescopes could end up beyond the reach of US astronomers

Nature, Comment, Matt Mountain and Adam Cohen


The science has never been so exciting. Earth-like planets have been found orbiting other stars2. Cosmologists are quantifying mysterious forces of ‘dark matter’ and ‘dark energy’3. Completely new windows have been opened onto the cosmos thanks to facilities such as the Atacama Large Millimeter/submillimeter Array (ALMA) in Chile and the Laser Interferometer Gravitational-Wave Observatory (LIGO), in the states of Washington and Louisiana.

But large facilities that can explore these frontiers cost billions of dollars and take decades to design, build and operate. ALMA was proposed in 1990 and became operational in 2013. The James Webb Space Telescope (JWST) was approved in 2000 and will be launched in 2021. The Large Synoptic Survey Telescope, recommended in 2010 and under construction in Chile, will begin to map the sky in 2023. The fruits of the 2020 Decadal Survey won’t see light until the 2030s.

The US community faces a daunting task. Each generation of facilities is getting more expensive and harder to build. Operational costs are mounting. Meanwhile, the research budgets of the US National Science Foundation (NSF) and NASA have remained more or less flat since the 1990s (see ‘Astronomical costs’). Hard decisions have been made to close old but still-productive telescopes, which has proved insufficient to pay for new ones. And these pressures will only get worse as more big projects come online.


Westat – Advancing Clinical Research



Rockville, MD September 12. Keynote: Atul Butte from University of California-San Francisco. [registration required]

HackGT 5

HackGT Team


Atlanta, GA October 19-21 at Georgia Tech. Deadline to register is September 22. Deadline for travel reimbursement applications is August 19.


Are you part of a team working on AI or ML products?

Please help [Hal Daume III and team] design tools to help create fairer AI/ML systems. “We’re conducting an anonymous survey (~20 min) to better understand teams’ practices, challenges, and needs around fairness & bias.”

Research Updates: Economic Graph Research Program CFP

“I am happy to provide an update on the new research teams that have joined the EGRP on an ongoing basis, and to announce a new call for proposals for new teams who are looking to pursue challenging questions in the fields of economics, analytics, or artificial intelligence, in partnership with LinkedIn. Teams who are selected to participate in the EGRP will have the opportunity to pursue their research alongside our experts in these fields and access de-identified or aggregate data from LinkedIn in a way that respects the privacy of LinkedIn members.” Deadline for proposals is December 1.
Tools & Resources

Developing Research Data Management Services and Support for Researchers: A Mixed Methods Study

Partnership: The Canadian Journal of Library and Information Practice and Research, Laure Perrier


This mixed method study determined the essential tools and services required for research data management to aid academic researchers in fulfilling emerging funding agency and journal requirements. Focus groups were conducted and a rating exercise was designed to rank potential services. Faculty conducting research at the University of Toronto were recruited; 28 researchers participated in four focus groups from June– August 2016. Two investigators independently coded the transcripts from the focus groups and identified four themes: 1) seamless infrastructure, 2) data security, 3) developing skills and knowledge, and 4) anxiety about releasing data. Researchers require assistance with the secure storage of data and favour tools that are easy to use. Increasing knowledge of best practices in research data management is necessary and can be supported by the library using multiple strategies. These findings help our library identify and prioritize tools and services in order to allocate resources in support of research data management on campus. [pdf download]

UX Design Guide for Data Scientists and AI Products

KDnuggets, Syed Sadat Nazrul


Realizing that there is a legitimate knowledge gap between UX Designers and Data Scientists, I have decided to attempt addressing the needs from the Data Scientist’s perspective.

Modeling and Simulation in Python

Green Tea Press, Allen B. Downey


Modeling and Simulation in Python is an introduction to physical modeling using a computational approach. It is organized in three parts:

  • The first part presents discrete models, including a bikeshare system and world population growth.
  • The second part introduces first-order systems, including models of infectious disease, thermal systems, and chemical kinetics.
  • The third part is about second-order systems, including mechanical systems like projectiles, celestial mechanics, and rotating rigid bodies.

  • Amazon Releases Cameras Recap API

    ProgrammableWeb, Patricio Robles


    Amazon has announced the release of a Cameras Recap API that gives apps the ability to play video feeds recorded by smart home cameras.”

    Adding a pipe operator to Python

    Hacker Noon, Bogdan Popa


    The other day, I was looking at some Python code that was made up of a sequence of deeply-nested function calls and a thought immediately got pushed up brain’s stack:

    Wouldn’t it be cool if Python had an operator akin to Elixir’s pipe operator?

    Quantifying memories: mapping urban perception

    Medium, Daniel Quercia


    Planners and social psychologists have suggested that the recognizability of the urban environment is linked to people’s socio-economic well-being. We build a web game that puts the recognizability of London’s streets to the test. It follows as closely as possible one experiment done by Stanley Milgram in 1972. The game picks up random locations from Google Street View and tests users to see if they can judge the location in terms of closest subway station, borough, or region. Each participant dedicates only few minutes to the task (as opposed to 90 minutes in Milgram’s). We collect data from 2,255 participants (one order of magnitude a larger sample) and build a recognizability map of Lon- don based on their responses. We find that some boroughs have little cognitive representation; that recognizability of an area is explained partly by its exposure to Flickr and Foursquare users and mostly by its exposure to subway passengers; and that areas with low recognizability do not fare any worse on the economic indicators of income, education, and employment, but they do significantly suffer from socialproblems of housing deprivation, poor living conditions, and crime.


    Full-time positions outside academia

    Senior UI Engineer, Notebooks Infrastructure

    Netflix; Los Gatos, CA

    Leave a Comment

    Your email address will not be published.