Friday, February 14, 2020

Money Walks

Another cool visualization - Neal Agarwal has created a site, Printing Money, that visualizes earnings at various hourly rates, revenue of major companies, and spending by organizations like NASA and the US Military.


If you scroll partway down, you can enter your own hourly rate and watch the money scroll by. Fair warning, watching the growth of the National Deficit race by at $125m per hour is dizzying.

You can check out Neal's other coding projects here. I particularly enjoyed his Life Stats project, where I learned that the moon has orbited 509 times in my lifetime, I've blinked about 300,607,247 times, and there are 3,120,013,847 more people on Earth than when I was born. And I'm currently playing with Where does the day go?, which asks you a series of questions to let you visualize how you spend your time each day.

Thursday, February 13, 2020

Visualizing the Tallest Building in Each State

Via Digg:
This data visualization, put together by takeasecond on Reddit, shows the tallest building in all 50 states in 2020. As the graph demonstrates, the current tallest building in America is New York's One World Trade Center at 1,776 feet tall. In contrast, the shortest building on the list is the Decker Towers in Vermont at just 124 feet tall.

Tuesday, January 28, 2020

On Presidential Approval Ratings and Opinions on Conviction

New research from Gallup suggests that job approval of Trump is strongly correlated with opinions on whether he should be convicted in his impeachment proceedings, a relationship also observed during Clinton's impeachment proceedings:
Americans' support for the Senate convicting Clinton in 1999 was much lower than current support for convicting Donald Trump. Gallup's Jan. 22-24, 1999, survey (one of a number we conducted while Clinton was on trial) found 33% of Americans in favor of Clinton being found guilty and removed from office, while 64% were against. Our latest survey on Trump shows 46% in favor of his conviction.

In the 1999 survey, Clinton's job approval rating was 69%, much higher than Trump's current 44% approval. So, the lower support for Clinton's conviction went hand in glove with his approval rating: 64% were against conviction compared with his 69% approval rating, and 33% were in favor of conviction juxtaposed against a 29% disapproval rating.
One interesting observation from the survey is that people place a higher bar on conviction than they do disapproval of Trump's job performance:
As noted, Americans' views on the impeachment of President Trump largely correlate with their overall views of the job he is doing as president. There is, however, a slight but evident deviation from a 100% relationship between Trump approval and views on the desired impeachment verdict.

Just 4% of Americans who approve of the job Trump is doing as president say he should be convicted, while a modestly higher 15% of those who disapprove of Trump say he should be acquitted. Apparently, there is a slightly higher bar for conviction among those who don't like Trump than there is for acquittal among those who do like him (although both of these positions are very much in the minority).
Research also suggests that Trump's current approval rating is almost the highest it's been in his presidency, but is still a lackluster performance when compared to post-World War II presidents in general (who historically average 53 percent):


Enough about politics. Here's a cute puppy (mine)!

Tuesday, January 14, 2020

Updates

New year, new job, new blog post describing it all. On January 6, I started working as a Data Analyst at the American Board of Medical Specialties, which oversees certification and maintenance of certification activities for 24 Member Boards (such as the American Board of Dermatology, American Board of Nuclear Medicine, and so on).

The main part of my job will be doing analysis, research, and program evaluation of the CertLink program, which is a really cool online system that tests physician knowledge in their certification area, provides feedback and introduces new information to improve over time, and measures the relevance of items to their practice, so that their maintenance of certification assessments can become more targeted to the population and types of cases they encounter in their practice. We're hoping that this kind of system will become the future of medical specialty certification, so rather than taking a high stakes exam every 10 years, medical specialists can maintain their certifications through targeted, longitudinal assessment and continuing education. And we're hoping to show this approach works by tying it to long-term, quality of care outcomes, like prescribing patterns. I'll share more as I learn more about the company and my role, to the degree that I can based on data privacy. But I'm so excited to be involved with this, using my psychometrics and statistics skills for the data I'm working with, and my research/program evaluation skills to show (how) the system works. I also finally get to use my SQL knowledge as part of my job, and will be using my R and Python programming skills pretty regularly as well.

Zeppelin is adjusting well to me working again. He adores his dog walker, who he sees three times a week, and has made many new friends in the doggy daycare he attends twice a week. He also has a huge crush on Mona, who can be found at Uncharted Books, stopping to stare longingly at her every time we walk by the shop. As is the case with so many crushes, this love seems to be unrequited; Mona tolerates Zeppelin but doesn't like the way he drinks out of her water bowl when we stop in.

On the blogging front, I'm working on an analysis of the 88 books I read last year, and might even do some long-term analysis of my last few years of reading data. Stay tuned for that.