Data Science newsletter – November 1, 2019

GROUP CURATION: N/A

This week Rock Health released a new report conducted in partnership with Stanford University that analyzes how the consumerization of healthcare is shaping the future of the industry. It broke down three major trends: how patient data is creating opportunities and challenges in the field, how online health information is changing the provider relationship and attitudes over a consumer’s willingness to share health data.

Researchers found that over the last two years, more patients are using digital health tracking tools, and in turn are sharing the data from those tools with their providers. Patients are also seeking health information online to inform discussions with their doctors — often in the form of self-diagnosing, coming up with treatment plans and asking to add or stop a medication.

… Data scientists from Northwestern University and the University of Chicago looked at the dynamics of failure in three different areas – science, entrepreneurship and terrorism – and found that the way one fails matters. The study, which will be published Oct. 30 in Nature, was led by a team at Northwestern’s Center for the Science of Science and Innovation (CSSI), Kellogg School of Management and McCormick School of Engineering.

After an initial failure, paths diverge, the study found. Some individuals go on to achieve eventual success, and others continue to fail until they drop out. This divergence was evident as early as the second attempt. The factor that ultimately determined which path an individual took was the extent to which they learned from previous failures and how they applied that knowledge going forward, according to the study.

“If you only look at the attributes of successful attempts, you’re missing half the story,” said corresponding author Dashun Wang.

There was significant variation in manuscript submission requirements for journals within the same scientific focus and only 4% of journals offered a fully format-free initial submission. Of 203 authors responding (71.5% response rate), only 11.8% expressed satisfaction with the resubmission process and 91% desired reforming the current system. Time spent on reformatting delays most publications by at least two weeks and by over three months in about 20% of manuscripts. The effort to comply with submission requirements has significant global economic burden, estimated at over $1.1 billion dollars annually when accounting for a research team’s time.

Interpretation

We demonstrate that there is significant resource utilization associated with resubmitting manuscripts, heretofore not properly quantified. The vast majority of authors are not satisfied with the current process. Addressing these issues by reconciling reformatting requirements among journals or adopting a universal format-free initial submission policy would help resolve a major subject for the scientific research community and provide more efficient dissemination of findings.

Advocates for genomic research in Africa are worried about fallout from a dispute that has roiled the Wellcome Sanger Institute, a major genome research center in Hinxton, U.K. Last year, whistleblowers privately accused Sanger of commercializing a gene chip without proper legal agreements with partner institutions and the consent of the hundreds of African people whose donated DNA was used to develop the chip. “What happened at Sanger was clearly unethical. Full stop,” says Jantina de Vries, a bioethicist at the University of Cape Town in South Africa, who has followed the dispute.

The National Science Foundation has awarded the University of Hawaiʻi’s Hawaiʻi Data Science Institute (HI-DSI) $1 million for the development of a web-based programming interface called Tapis, in partnership with the Texas Advanced Computing Center (TACC) and the University of Texas at Austin. The development of Tapis will provide scientists with important tools to gather data and conduct computationally intensive research.

The framework, which will serve a diverse group of users, can help automate and streamline large workflows or pipelines of software applications and allow scientists easier, user-friendly access to computational resources.

In some ways, drilling into Antarctica’s ancient ice is easier than interpreting it. Today, more than 2 years after presenting the discovery of the world’s oldest ice core, scientists have published an analysis of the 2.7-million-year-old sample. One surprising finding: Air bubbles from 1.5 million years ago—from a time before the planet’s ice age cycles suddenly doubled in length—contain lower than expected levels of carbon dioxide (CO2), a possible clue to the shift in the ice age cycle.

The CO2 levels are “amazingly low,” says Yige Zhang, a paleoclimatologist at Texas A&M University in College Station. He adds that the study, published today in Nature, is “quite interesting” because it reports the first direct measurements of atmospheric gases from that mysterious time.

Analytics at Wharton has selected eight projects as inaugural recipients of the newly created Data Science and Business Analytics Fund. The projects, chosen from 23 proposals submitted from around Wharton, will focus on a range of topics and include: predictive analytics to detect financial irregularities, machine learning as a tool for human resources and the creation of a new Women in Analytics conference. A formal call for submissions solicited proposals that demonstrate the need for financial support and infrastructure to enhance faculty research, student learning opportunities and engagement with industry and alumni.

“We are delighted by the quantity and caliber of the proposals submitted to Analytics at Wharton’s Data Science and Business Analytics Fund,” said Vice Dean of Analytics, Eric T. Bradlow, a professor of marketing, economics, education and statistics and the chair of Wharton’s Marketing Department. “All of the proposals were intriguing in their own way and collectively reflect the incredible breadth of analytics knowledge being created at Wharton.”

One of the guests at the announcement Wednesday of the University of Michigan Innovation Center in Detroit gushed to me that the U-M project was the biggest thing for Detroit since the assembly line.

Hyperbole? Perhaps. But maybe not so far wrong if the U-M project delivers everything that Mayor Mike Duggan and billionaire backers Dan Gilbert and Stephen Ross promise.

When the initial piece opens in about four years, U-M’s Detroit Center for Innovation will create a mini-campus for about 1,000 U-M graduate students working in tech fields like artificial intelligence, mobility, data science and cybersecurity.

That will prove so significant, say the project’s backers and creators, that it will bolster Detroit’s claim to be an international hub of tech innovation.

When people talk about AI, Jason Harper says they often focus on “autonomous vehicles or image recognition or other things that would be bleeding-edge AI technology research.”

Harper, founder of the Ann Arbor cloud-based software company RXA, says those discussions are “great and exciting.” But they overlook the many smaller problems those in the AI field are solving – and advancements they’re making – right now. That’s why RXA and Utah-based Domo created a one-day conference called A2.AI, held at SPARK Central in downtown Ann Arbor on Oct. 25.

Carnegie Mellon University and Northrop Grumman Systems Corporation entered a new five-year master agreement to expedite funding of science and engineering research projects at the university.

The agreement expands and strengthens the long-standing relationship between the two instiutions. For years, Northrop Grumman has supported research in various areas within CMU, including projects at its Entertainment Technology Center, Software Engineering Institute and CyLab, among others.

Virginia Tech is accepting applications for the first degree program to be run out of its new Innovation Campus at Potomac Yard in Alexandria.

The program is a new master’s of engineering degree in computer science that was just approved by the State Council of Higher Education for Virginia. The university is now taking applications for spring 2020 enrollment — due Nov. 15 — and fall 2020 enrollment, which has a deadline of Feb. 1.

The master’s program will be offered part-time, full-time or full-time accelerated. The degree requires 30 hours of graded coursework; Full-time students should be able to complete the degree in three semesters, according to a press release from Virginia Tech.

Scientists from Intel, Amazon, Facebook and Google joined some of the leading academic minds on artificial intelligence (AI) to discuss the future of machine learning during the inaugural Responsible Machine Learning Summit hosted by UC Santa Barbara. With more than 120 students, faculty, business leaders and invited guests on hand, every speaker and panelist agreed on the importance of establishing an ethical foundation for machine learning, in which a computer uses algorithms and data to make predictions or decisions on its own.

“We need to better understand the mutual influence between society and machine learning,” said William Wang, a professor of computer science and organizer of the event. “Personally, I’m interested in improving the quality of life by learning the important societal factors and impacts that should be considered when building algorithms, such as fairness, transparency, privacy and accountability.”

“And don’t forget to download our app!” the lady on reception told my partner and I as we dropped our baby off at nursery for the first time. I nodded obligingly as we walked out the door. Minutes later, both our phones pinged informing us that we had been given access to parental accounts that would allow us to monitor our baby. So began a daily ritual of checking in on how many times said baby had pooped.

This, right here, is technological progress. We can, with just a couple of taps, check how many bowel movements our baby has had and at what time, find out how much of their lunch they have eaten and when and for how long they have napped. There’s even a chart for tracking the length of their naps over time. It’s both glorious and terrifying.

The app in question, Famly, is the work of an eponymous Copenhagen-based startup, which has to date raised more than £322,000 in seed funding. Famly sits at the more sensible end of an ever-growing industry of products and services that aim to quantify our babies. By 2024, the global interactive baby monitor market is expected to top $2.5 billion (£1.93bn). And today, nobody stops with the purchase of a rudimentary baby monitor.

The software giant Microsoft plans to open a hub in Syracuse aimed at developing uses for new technology, aiding tech start-ups and training the workforce of the future.

Microsoft will establish a “Smart Cities Technology” hub as part of Mayor Ben Walsh’s Syracuse Surge plan to bolster the regional economy with high-tech jobs. It’s Microsoft’s third such hub in the nation and the first in the Northeast.

A pay gap between female and male employees at US federal science agencies decreased but persisted over 15 years, and the source of the disparity wasn’t the same across the board, according to a new study (Am. J. Sociol. 2019, DOI: 10.1086/705514).

An academic research team looked at the pay and position of federal employees, including almost 2.8 million staff members at seven science agencies, from 1994 to 2008. All agencies saw a substantial decrease in the pay gap during that time period. At the National Science Foundation, for example, women made 58 cents for every dollar made by men in 1994. By 2007–8, that had risen to 73 cents.

When the researchers looked for the source of the pay gap, they found that it was different at different agencies. Physical science-focused agencies—the Department of Energy and National Oceanic and Atmospheric Administration—had the largest pay gaps between men and women working in the same jobs. Whereas at life-science and interdisciplinary agencies—the NSF, National Institutes of Health, Department of Agriculture, Centers for Disease Control and Prevention, and Environmental Protection Agency—the pay gap was caused by hiring more women into lower level positions rather than paying them less than men in the same position.

Stanford, CA November 14, starting at 5:30 p.m. Panelists are Rene Almeling
and R. Alta Charo, moderated by Debra Satz. [registration required]

New York, NY November 12, starting at 9 a.m., Ford Foundation Center for Social Justice (320 East 43rd St.) “A convening exploring the intersection of big data, gender equality, and decision making.” [free, registration required]

You might think the first thing data scientists do on a project is grab a bucket of data and run with it. And for good reason—that’s how many data scientists write about projects after the fact. But in practice, when we’re making something genuinely new, we often don’t have data to begin with, nor do we even know what kind of data we will need. We’re staring at a blank canvas, so the only way to dive in is to listen, observe, and sketch.