Roger Peng outlines four main roles of a data scientist:

If you’re reading this and find yourself saying “I’m not an X” where X is either scientist, statistician, systems engineer, or politician, then chances are that is where you are weak at data science. I think a good data scientist has to have some skill in each of these domains in order to be able to complete the basic data analytic iteration.

The good thing about data science is that you can apply the skills to different fields and tasks. It’s also one of the challenges when you’re in the early phases of learning, because you have to figure out what to work on. This should point you in the right direction.

See also: Peng’s tentpoles of data science.

As we’ve talked about before, it can be hard to really understand the scale of big numbers. So when we hear that over 250,000 people died because of the coronavirus, it can be hard to conceptualize that number in our head. Lauren Tierney and Tim Meko for The Washington Post provide a point of comparison by highlighting counties that have have populations under 250,000.

Whole counties, or whole clusters of counties, that would be completely wiped out.

It’s a lot.

It’s grown easier and easier to generate fake faces with AI. For The New York Times, Kashmir Hill and Jeremy White demonstrate the tech with a slick interactive. Quickly adjust age, eye, mood, and gender. All fake.

It was only a few years ago when the idea seemed novel. One year later, there were guides (and warnings) for spotting fake faces. By 2019, there was a marketplace for fake faces (of course). Sometimes it’s scary to think about what the internet will be in five years.

In any case, check out the NYT piece. The smooth transitions between faces, one facial aspect at a time, is mesmerizing.

Reporting for NPR, Sean McMinn and Selena Simmons-Duffins on staffing shortages:

On data availability:

This is the first time the federal agency has released this data, which includes limited reports going back to summer. The federal government consistently started collecting this data in July. After months of steadily trending upward, the number of hospitals reporting shortages crossed 1,000 this month and has stayed above since.

The data, however, are still incomplete. Not all hospitals that report daily status COVID-19 updates to HHS are reporting their staffing situations, so it’s impossible to tell for sure how much these numbers have increased.

The first time.

It was back in March, a few lifetimes ago, when we were talking about flattening the curve so that hospitals could provide care to those who needed it. This federal dataset is just coming out now in November? Obscene.

A small gathering of 10 people or fewer can seem like a low-risk activity, and at the individual level, it’s lower risk than going to a big birthday party. But when a lot of people everywhere are gathering, small or large, the collective risk goes up. For FiveThirtyEight, Maggie Koerth and Elena Mejía illustrate the reasoning.

The collective part is where many seem to get tripped up. “Flattening the curve” only works when everyone works together. Lower your risk, and you lower the collective risk. You’re helping others. You’re helping those you care about.

Then, collectively, we all get out of this mess.

For The Pudding, Michelle McGhee analyzed representation in crossword puzzles. Some crossword publications do better than others.

As of December 2019, The USA Today puzzle is edited by Erik Agard, a 27-year old crossword champ who told me, “bringing some balance on the representation front is something I actively try to do.” A prominent crossword blogger called USA Today’s puzzle “the most interesting, innovative, and provocative daily crossword” out right now. Let’s take a look at how USA Today, and other publications, are taking a puzzle that’s been called too old, too white, too male, and changing it up.

The story also comes with playable, data-generated puzzles so that you can feel the difference over decades.

Here’s the good stuff for November.

The University of Oxford’s Blavatnik School of Government defined an index to track containment measures for the coronavirus. For The New York Times, Lauren Leatherby and Rich Harris plotted the index against cases and hospitalizations:

When cases first peaked in the United States in the spring, there was no clear correlation between containment strategies and case counts, because most states enacted similar lockdown policies at the same time. And in New York and some other states, “those lockdowns came too late to prevent a big outbreak, because that’s where the virus hit first,” said Thomas Hale, associate professor of global public policy at the Blavatnik School of Government, who leads the Oxford tracking effort.

A relationship between policies and the outbreak’s severity has become more clear as the pandemic has progressed.

States with more restrictions tend to have lower rates.

From these plots, it seems clear what we need to do. But I think most people have made up their minds already, and the interpretation of the data leads people to different conclusions.

With the holidays coming up, I just hope you lean towards clarity.

For The New York Times, Ford Fessenden, Lazaro Gamio and Rich Harris go with a Dorling cartogram to look at the votes gained per county in the 2020 election, compared against the 2016 election.

As you’d expect, voting overall was up just about everywhere this year. Some counties shifted left. Some shifted right. The key points of interest come about when the the map starts zooming into specific regions.

See also: the election wind map.

Alan McConchie from Stamen recaps the wide array of maps and charts that came out before, during, and after election night:

This year we saw continued refinement of traditional election maps styles, and an exciting (and nerve-wracking) new frontier developed with the visualization of post-election ballot counting. Dataviz practitioners are struggling with challenges of how to show uncertainty and how much uncertainty can be shown while still making our visualizations clean and easy to understand. Election cartographers are dealing with their own dilemma of how much to show the polarization and inequality that currently exists in our electoral system (with the risk of reinforcing it) versus making counterfactual maps of systems that could or should be.

[via Co.Design]

Joseph Cox, reporting for Motherboard:

Some app developers Motherboard spoke to were not aware who their users’ location data ends up with, and even if a user examines an app’s privacy policy, they may not ultimately realize how many different industries, companies, or government agencies are buying some of their most sensitive data. U.S. law enforcement purchase of such information has raised questions about authorities buying their way to location data that may ordinarily require a warrant to access. But the USSOCOM contract and additional reporting is the first evidence that U.S. location data purchases have extended from law enforcement to military agencies.

Oh.

The World Bank tracks global development through a number of indicators. (You can see and download much of the data through their catalog.) With a story-based approach, they published an atlas for 2020 that focuses on 17 development goals, such as end poverty, end hunger, and stop global warming. There’s one story per goal, charting out multiple indicators in each story.

There’s a lot to look at, but one thing you’ll probably notice across all of the topics is progress. It’s not all spikes and waves out there.

Voter turnout this election was higher than it’s been in a long time, but the winner margins were still small. Alyssa Fowers, Atthar Mirza and Armand Emamdjomeh for The Washington Post showed the margins with dots. Each circle represents 3,000 votes, and the blue and red circles represent by how much the candidate won by in a given state.

The dots showing absolute counts are useful to see the scale of each win, which percentages don’t capture.

There’s a video (one of too many I am sure) going around that “shows” election rigging. Statistician Kristian Lum shows, with good ol’ basic math and R plots, why the “evidence” is what happens during a normal election.

[arve url=”https://www.youtube.com/watch?v=fH0Z-8563-E” loop=”no” muted=”no” /]

Coronavirus cases are rising (again), which includes prisoners and prison staff. The Marshall Project has been tracking cases since March and provides a state-by-state rundown:

New infections this week rose sharply to their highest level since the start of the pandemic, far outpacing the previous peak in early August. Iowa, Michigan and the federal prison system each saw more than 1,000 prisoners test positive this week, while Texas prisons surpassed 2,000 new cases.

To combat cheating during online exams, many schools have utilized services that try to detect unusual behavior through webcam video. As with most automated surveillance systems, there are some issues. For The Washington Post, Drew Harwell looks into the social implications of student surveillance:

Fear of setting off the systems’ alarms has led students to contort themselves in unsettling ways. Students with dark skin have shined bright lights at their face, worrying the systems wouldn’t recognize them. Other students have resorted to throwing up in trash cans.

Some law students who took New York’s first online bar exam last month, a 90-minute test proctored by the company ExamSoft, said they had urinated in their chairs because they weren’t allowed to leave their computers, according to a survey by two New York state lawmakers pushing to change the rules for licensing new attorneys during the pandemic.

Oh.

When every day feels like more of the same, here are some of the things I do to shake off the monotony.

For The New York Times, Denise Lu and Karen Yourish looked at the red and blue shifts for the counties that voted red in 2016:

President-elect Joseph R. Biden Jr. won the popular vote by more than five million — and his margin is expected to grow as states finish counting. Still, results so far show that President Trump’s support remained strong in most of the counties that voted for him in 2016. Here’s how.

Always enjoy scrollytelling through spaghetti.