Data Science newsletter – April 7, 2020

GROUP CURATION: N/A

With articles about “silent spreaders” and “stealth transmission” flying across the internet, friends were starting to text me: Was it still OK to go for a walk with a friend, even 6 feet apart? Or should all interaction be avoided? Should we start wearing masks to the grocery store? At the same time, my colleagues were scrutinizing guidelines at various workplaces and agencies we cover: The New York City Fire Department told workers on March 19 they were to come to work, so long as they had no symptoms, even if they had had “close contact with someone who is a known positive COVID-19 patient,” according to a document obtained by ProPublica. Was that policy wise?

I decided to dive into the available data. What I discovered is that not only can people be infected and experience no symptoms or very mild symptoms for the first few days, but this coincides with when the so-called viral load — the amount of virus being emitted from an infected person’s cells — may be the highest. That makes the virus a truly formidable opponent in our densely packed, globally connected world. We’re going to have to be smarter than this virus to stay on top of it.

The coronavirus pandemic has stalled scientific fieldwork and may even start to affect the monitoring of the climate, scientists have warned.

Major projects to gather environmental data have been postponed or canceled over concerns that teams of researchers working together will spread the Covid-19 virus.

The crisis has so far mainly stymied long-term studies, but concerns have been raised that routine monitoring of weather and the climate crisis may be affected if the pandemic drags on for an extended period.

Petteri Taalas, secretary general of World Meteorological Organization, said: “The impacts of climate change and growing amount of weather-related disasters continue.

After much delay, the New York City Department of Health recently released data on the number of tests given and the number of positive COVID-tests in each ZIP Code. And what followed was a flurry of maps and analysis by news organizations hoping to pinpoint which neighborhoods were hardest hit.

I recently pointed out that subway ridership had fallen much slower in low income neighborhoods than in higher income ones, leading me to believe that the effect of COVID may be harsher in those neighborhoods as more people continued to venture out as things worsened (likely due to the nature of their work).

A quick glance at this new dataset would at first suggest otherwise.

In a highly anticipated decision, a federal judge in Washington, D.C. has ruled that violating a website’s terms of service does not violate the Computer Fraud and Abuse Act (CFAA), the notoriously vague federal computer crime law. This decision solidifies the trend in courts to reject overbroad and constitutionally troublesome interpretations of the CFAA that have threatened to criminalize innocuous online activity, such as beneficial research and journalism.

The case, Sandvig v. Barr, is a lawsuit by university professors, computer scientists, and journalists who want to research how algorithms unlawfully discriminate based on characteristics like race or gender. To determine whether algorithms produce discriminatory results, the researchers want to create multiple “tester” accounts. For example, to study the impact of gender on an employment website’s algorithms, a researcher may create tester accounts that are identical except as to their listed gender, to isolate the effect that users’ gender has on the job postings that an algorithm provides them.

WHOOP, a fitness band worn by both professional athletes and everyday people to track health and fitness levels, may be able to provide an early indicator of COVID-19 symptoms and have partnered with the Cleveland Clinic to test the theory.

One of the apparent underlying symptoms of the novel coronavirus is an elevation in respiratory rate, a measurement the WHOOP bands provide to its wearers.

“We are really pumped to have them as partners in this collaboration to better understand COVID-19,” said Will Ahmed, WHOOP CEO. “And you know, broadly speaking, it’s what everyone should be doing is trying to see what they can do to help understand this thing.”

The use of artificial intelligence (AI), technologies that can interact with the environment and simulate human intelligence, has the potential to significantly change the way we work. Successfully integrating AI into organizations depends on workers’ level of trust in the technology. A new review examined two decades of research on how people develop trust in AI. The authors concluded that the way AI is represented, or “embodied,” and AI’s capabilities contribute to developing trust. They also proposed a framework that addresses the elements that shape users’ cognitive and emotional trust in AI, which can help organizations that use it.

The review, by researchers at Carnegie Mellon University and Bar Ilan University, appears in Academy of Management Annals.

“The trust that users develop in AI will be central to determining its role in organizations,” explains Anita Williams Woolley, Associate Professor of Organizational Behavior and Theory at Carnegie Mellon University’s Tepper School of Business, who coauthored the study. “We addressed the dynamic nature of trust by exploring how trust develops for people interacting with different representations of AI (e.g., robots, virtual agents, or embedded) as well as the features of AI that facilitate the development of trust.”

In this new world requiring physical distancing, how can we temporarily redesign our public physical environments to make it easier for us to stay safely away from each other? Such design is important because our habits are triggered by cues built into our physical environment. If everything looks and feels the same, we will habitually revert to old behaviors.

We may remember to keep a safe distance sometimes, but we will fail a lot, which has serious consequences in this pandemic. What’s worse is that we won’t realize how much we’re failing because we will be acting unconsciously. If we can design quick, temporary solutions to alter these physical cues, we can trigger better, more effective physical distancing behavior, and in turn help #FlattenTheCurve.

How We Feel, a crowdsourced COVID-19 symptom tracking app that respects privacy, launched Thursday to provide more data on the spread of the coronavirus.

Pinterest Inc. Chief Executive Ben Silbermann partnered with researchers from Harvard University, Stanford University and Feeding America to produce the app, now available for iOS and Android devices. It asks people to take 30 seconds each day to submit information on how they feel – healthy or otherwise. Other data that the app collects includes age and ZIP code, but no other information is requested or collected.

Users can opt to also share their symptoms if they are sick and also information on what they are doing to prevent the spread of the disease, such as social distancing or self-isolation. Pinterest said the data will only be used in support of research tracking the spread of SARS-CoV-2, the coronavirus that causes the disease COVID-19.

Amid the current need to continually focus on the COVID-19 crisis, it is understandably hard to address other important issues. But, this morning, Washington Governor Jay Inslee has signed landmark facial recognition legislation that the state legislature passed on March 12, less than three weeks, but seemingly an era, ago. Nonetheless, it’s worth taking a moment to reflect on the importance of this step. This legislation represents a significant breakthrough – the first time a state or nation has passed a new law devoted exclusively to putting guardrails in place for the use of facial recognition technology.

Without social distancing, one coronavirus sufferer could, in six weeks, have started chains of transmission with 1,093 cases;

With social distancing, the same sufferer could, in six weeks have started chains of transmission with 127 cases.

As part of the war on coronavirus, U.S. regulators are taking aggressive steps against “sludge” – paperwork burdens and bureaucratic obstacles. This new battle front is aimed at eliminating frictions, or administrative barriers, that have been badly hurting doctors, nurses, hospitals, patients, and beneficiaries of essential public and private programs.

Increasingly used in behavioral science, the term sludge refers to everything from form-filling requirements to time spent waiting in line to rules mandating in-person interviews imposed by both private and public sectors. Sometimes those burdens are justified – as, for example, when the Social Security Administration takes steps to ensure that those who receive benefits actually qualify for them. But far too often, sludge is imposed with little thought about its potentially devastating impact.

The coronavirus pandemic is concentrating the bureaucratic mind – and leading to impressive and brisk reforms.

The Centers for Disease Control and Prevention has begun preliminary studies to try to determine how many Americans have already been infected with SARS-CoV-2, the virus that causes Covid-19, an agency official revealed Saturday. On Friday, the agency said nearly 240,000 people in the country have been infected with the virus and nearly 5,500 have died.

Joe Bresee, deputy incident manager for the CDC’s pandemic response, said the agency hopes to flesh out the portion of cases that have evaded detection using three related studies.

The first, which has already begun, will be looking at blood samples from people never diagnosed as a case in some of the nation’s Covid-19 hot spots, to see how widely the virus circulated. Later, a national survey, using samples from different parts of the country, will be conducted. A third will look at special populations — health care workers are a top priority — to see how widely the virus has spread within them.

Many of us know the frustration of clicking on a news article and running into a paywall. Academic researchers also run into this problem when trying to access journal articles, but for them, the cost can often run into the thousands.

A new open-access agreement between Caltech and the Association for Computing Machinery (ACM), an academic society for computing research, guarantees that all papers authored by Caltech researchers that appear in ACM journals will be freely accessible to any user without cost.

“This is a big deal for Caltech because of the impact our research has worldwide,” says University Librarian Kara Whatley. “Many researchers are at institutions that can’t afford to buy subscriptions to the journals our authors publish in, which effectively hides Caltech research from them.”

When an online conversation derails, participants might look to artificial intelligence to help them get back on track, Cornell University research shows.

Humans having difficult conversations said they trusted artificial intelligence systems — the “smart” reply suggestions in texts — more than the people they were talking to, according to the study, “AI as a Moral Crumple Zone: The Effects of Mediated AI Communication on Attribution and Trust,” published online in the journal Computers in Human Behavior.

The National Science Foundation-funded research has new relevance, with most daily conversations taking place online because of social distancing.

“We find that when things go wrong, people take the responsibility that would otherwise have been designated to their human partner, and designate some of that to the artificial intelligence system,” said Jess Hohenstein, an information scientist and the paper’s first author. “This introduces a potential to take AI and use it as a mediator in our conversations.”

HHS won’t enforce penalties for violations of certain provisions of the HIPAA privacy rule against healthcare providers or their business associates for good-faith disclosures of protected health information for public health purposes during the COVID-19 emergency.

The HHS Office for Civil Rights said Thursday that it was exercising its enforcement discrimination in making the policy change during the declared emergency period. The notification was issued to support federal and state agencies, including the CMS and the Centers for Disease Control and Prevention, that need access to COVID-19 related data including protected health information.

Online April 25-26. “If you or your company wants to present data effectively using data visualization, needs in-depth background knowledge about the art and science of data visualization, or needs to acquire data visualization literacy, this event is for you!” [$$$, limited to 30 attendees]

Online April 15, starting at 5:30 p.m. EDT. “In this session, chaired by Bruno Dupire, Alexander Lipton will present his current research, followed by several ‘lightning talks’ of 5 minutes each in quick succession. This format gives the audience the opportunity to be exposed to a wider variety of topics.” [registration required]

The up and coming JavaScript framework Svelte, created by visual journalist and software developer Rich Harris, aims to make it easier to write faster, smaller interactive websites and applications. Web developer Shawn Wang says he cut the size of his personal website from 187 kilobytes to 9 kilobytes by switching from React to Svelte.

“It was a big ‘wow’ moment,” Wang says. “I wasn’t even trying to optimize for size and it just dropped.”

In our ACM CHI Conference on Human Factors in Computing Systems (CHI 2020) paper “Optimizing for Happiness and Productivity: Modeling Opportune Moments for Transitions and Breaks at Work,” we—along with PhD students Harmanpreet Kaur and Alex Williams, both Microsoft interns at the time—investigated how to optimize happiness (as quantified via facial expressions) and productivity by modeling information workers’ affect and work patterns. The result is a group of models designed to capture when a person should consider switching to a new task or stepping away for a break as a means of enhancing positive emotions while still ensuring work gets done.

We offer methods to analyze the “differentially private” Facebook URLs Dataset which, at over 10 trillion cell values, is one of the largest social science research datasets ever constructed. The version of differential privacy used in the URLs dataset has specially calibrated random noise added, which provides mathematical guarantees for the privacy of individual research subjects while still making it possible to learn about aggregate patterns of interest to social scientists. Unfortunately, random noise creates measurement error which induces statistical bias — including attenuation, exaggeration, switched signs, or incorrect uncertainty estimates. We adapt methods developed to correct for naturally occurring measurement error, with special attention to computational efficiency for large datasets. The result is statistically consistent and approximately unbiased regression estimates and descriptive statistics that can be interpreted as ordinary analyses of non-confidential data but with appropriately larger standard errors.