NYU Data Science newsletter – August 3, 2016

GROUP CURATION: N/A

For the second year in a row, WrangleConf did not disappoint. The conversation picked up right where last year’s left off: on the ethics of our craft. Last year the focus was on the humans building algorithms and the humans whom algorithms affect. This year, the discussion expanded in scope to consider the growing number people who interact with data science teams.

With an eye toward the increasing presence of data science in our daily lives, the speakers were more focused than ever on strategies to build and maintain trust: opening communication, recognizing bias, and, well, giving a damn.

There’s a stereotypical image of a librarian in popular culture: someone older, in thick-rimmed glasses and overly modest clothing, guarding the silence in a room full of books with all-powerful shushes.

But as the internet has largely replaced brick-and-mortar libraries as the go-to resource for information gathering, librarians’ purview is no longer confined to just books. Libraries have had to evolve from providing the internet as a service, to being responsible for interacting with it, to indexing and archiving a rapidly increasing amount of information. Though the occupation is only expected to grow by 2 percent from 2014 to 2024, many librarians have forgone bookkeeping and cataloging for specializing in multimedia and taking on research- and technology-oriented projects such as digitizing archives.

Theresa Quill, a research librarian at Indiana University, Bloomington, specializes in the relationship between geography and cultural behavior, and digital mapping. While she assists students in the same ways librarians traditionally have, she also works on projects like making maps based on interesting novels and indexing Russian war maps.

The National Science Foundation (NSF) announced two major awards to establish Scientific Software Innovation Institutes (S2I2). The awards, totaling $35 million over 5 years, will support the Molecular Sciences Software Institute [led by Virginia Tech] and the Science Gateways Community Institute [led by University of California-San Diego], both of which will serve as long-term hubs for scientific software development, maintenance and education.

The city has made a commitment to Vision Zero, a global movement to reduce traffic fatalities and severe injuries to zero, and is now looking to determine the most effective policies and interventions to make streets safer. Building off our recent work with New York City in partnership with Microsoft Tech and Civic Engagement, we’re conducting an in-depth pedestrian and bicyclist safety study for the City of Seattle to empower local decision makers with potentially life-saving information.

To lay the groundwork for this study, over 20 data scientists volunteered at a DataDive in May sponsored by Microsoft and hosted by University of Washington’s eScience Institute.

NLP technology is advancing rapidly as the Web continues its explosive growth. An abundance of electronic text and spoken-text data is making it easier for computer scientists to train their algorithms to do more ambitious work. Much of the new technology that Columbia researchers will unveil Aug. 7-12, at the Association for Computational Linguistics’ annual meeting in Berlin, makes creative use of the proliferating text data around us. Below is a summary of the papers they will present.

Kimera Systems, developers of the commercial human-like intelligence technology for connected devices, announced the closing of its angel round of funding. The funding, received from private investors, allows the company to begin scaling its breakthrough Nigel™ artificial general intelligence (or AGI) technology and bring it to market.

Nintendo’s Pokémon GO has already overtaken Candy Crush Saga to become one of the most-used apps. One of the main draws of the game for many people is that it gets players off their couches and out into the real world in search of Pokémon, and this post by Lynne Slowey on IBM’s Internet of Things blog highlights an impressive use of the Watson API to help players find these virtual creatures.

This year we’ve seen a spate of announcements from Google to Facebook to IBM and scores of smaller companies, announcing Artificial Intelligence, or “AI” initiatives. Fueled by technology advancements (big data processing power), IBM Watson’s marketing and sales engine and Facebook’s marketing engine, media are latching onto AI as the next big technology trend. At first, it seemed like 2016 was the year that that this technology finally arrived, but the reality is that today’s AI solutions are still critically dependent on humans.

In the past five years, fueled by about $30 billion in federal incentive payments, medicine has finally become a digital industry. More than 90% of American hospitals now have electronic health records, as do the vast majority of physician offices. Decades after most other information-intensive industries switched from paper to silicon, in medicine, the x-rays, the three-ring binders, and the card tables have finally left the building.

Clearly, the world of today’s physicians will be vastly different from the world I entered in the early 1980s. Just as clearly, the training of future physicians must evolve for their work in a digital healthcare system. But how should it change?

A scientific expedition into the DNA of more than 450,000 customers of gene-testing company 23andMe has uncovered the first major trove of genetic clues to the cause of depression.

The study, the largest of its kind, detected 15 regions of human genome linked to a higher risk of struggling with serious depression. The study was carried out by drug giant Pfizer as part of an alliance with 23andMe, the California company whose gene reports have been purchased by more than 1.2 million people (see “50 Smartest Companies 2016: 23andMe”).

A review of hundreds of the candidates’ messages on Twitter and Facebook—using data* from Illuminating 2016, a project supported by the Tow Center for Digital Journalism and Syracuse University’s Center for Computational and Data Sciences—suggests the candidates took very different approaches to the digital medium in the month before Clinton clinched the nomination.

Sanders’s feed had fewer negative messages, relying more on calls to action than Clinton’s feed. And despite his success with small donations, Sanders was more focused on getting out the vote than urging supporters to donate money on social media. Clinton’s strategic use of calls to action was focused on digital engagement—and attempts to create an open and collaborative campaign environment by inviting supporters to engage with policy discussions online.

Kay, 76, is now the President of the Viewpoints Research Institute, after four decades of computing work at PARC, Apple, Disney and HP. In an interview with FactorDaily over three emails — we believe it is the first he’s given to an Indian publisher — Kay, talks about today’s computing challenges, artificial intelligence, learnings for modern research, and, of course, Apple.

It might feel like using data is big news now, but the truth is that we’ve been using data for a long time in the Internet business. For the past 20 years, we’ve been moving and replicating more and more experiences that we used to have in the physical world into the digital world. Sharing photos, having conversations, duties that we used to perform in our daily work have all become digital. We could probably have a separate discussion as to how much the migration from the physical “real” world to the digital world has benefitted or been detrimental to our society, but you can’t deny that it’s happening and only continues to accelerate at a breakneck pace.

Let’s take a look at what it means for these experiences to be moving from the physical to the digital.

Working remotely sounds great. We think we’ll save ourselves the commute and we’ll be able to flexibly weave in and out of work and home life. In reality, work takes a different shape when there is no office, and we’re all in different environments trying to connect to other humans.

We have no “central” office at Hypothesis and we currently work from 4 different countries in different time zones. We have learned that communication, decision making, social interaction and leadership need slightly different strategies. We still have much to improve, but we’re working on it every day.

Slide 1:

D3 v3 was one large library. You had to include all of it, even if you didn’t need all of it.

D3 v4 is actually a set of small modules. You can choose to use one or more as you need them.

Google Cloud Machine Learning is a managed platform that enables you to easily build machine learning models, that work on any type of data, of any size. Create your model with the powerful TensorFlow framework, that powers many Google products from Google Photos, to Google Cloud Speech. Build models of any size with our managed scalable infrastructure, which is powered by GPUs. Your trained model is immediately available for use with our global prediction platform that can support thousands of users and TBs of data. The platform is integrated with Google Cloud Dataflow for pre-processing, allowing you to access data from Google Cloud Storage, Google BigQuery, and others.