Tag: Data

This American Life – Fan Data and Insights

I was first introduced to This American Life about a decade ago, right after the podcast, then a relatively novel media format, got sensationalized by the true crime show Serial. Only later did I find out about the shared DNA between the two shows. Life comes full circle.

Over the past thirty years, since Ira Glass started This American Life, the show has grown from a personal passion project (formerly called Your Radio Playhouse) discussing the everyday lives of ordinary Americans into one that touches millions of listeners, covering major events across the globe and making a real impact on the people its team cares about.

As the show approaches its 900th episode, every one of which I’ve listened to at least once and many more than that, I decided to dissect it and find the recipe for what makes it such a fascinating show to me and millions of others: what it talks about week over week, who has been telling these stories, what makes them so immersive to listen to, and all the real-life impact it has made. Let’s go.

https://milesmeters.com/pieces/this-american-life/

I used 3 weeks to gather data, analyze, and build a site dissecting ~900 episodes of This American Life over 30 years. As a die-hard fan of the show, some insights are as expected (e.g., themes, contributors), but some are genuinely surprising (e.g., songs & artists).

https://www.reddit.com/r/ThisAmericanLife/comments/1wk1scu/900_episodes_over_30_years_data_insights_of_this/

Three Ways to Answer a Question

FISHER: Over 50 percent of Americans do want a partner who shares their political views. About 43 percent want a partner who is of the same ethnic background. About 46 percent want somebody of the same religious background. What’s interesting to me is the huge percentage of people that don’t care.

DUBNER: Is it that they don’t care, or they say on a survey they don’t care because they may want to appear to be the type of person who would say that they don’t care when, in fact, they may care?

FISHER: You never know, Stephen. I do a lot of questionnaires and you can answer a questionnaire in one of three ways:
with who you really are,
with who you want to be,
or with who you want others to think you are.
But because we have so many thousands of people, and there’s a bell-shaped curve, we can be pretty confident of what we’re doing.

Why Did You Marry That Person?
Sure, you were “in love.” But economists — using evidence from Bridgerton to Tinder — point to what’s called “assortative mating.” And it has some unpleasant consequences for society.

Forecasts are fundamentally partisan – Steve Ballmer Quote

On how USAFacts shares information during a time where many aren’t trustworthy of facts that don’t line up with their opinions

“We do a couple of different things. Number one, we only use government data and we only use data about history. Now, some people would say, ‘Hey, I don’t agree with government data.’ It’s just the best we have. And if citizens don’t think they can rely on government data, then it’s incumbent on us, all of us, to push our government for better, more consistent and more complete data. And we certainly see opportunities to do that. But by and large, we believe that our government’s statistical agencies have good people who are doing good work.

“The thing that we don’t do with that is try to do forecasts. Forecasts are fundamentally partisan. And I say that only in the sense that for every bright, smart economist that will say ‘X is going to happen,’ there’ll be another bright, smart economist that will say ‘not X is going to happen.’ So we think the thing that citizens need to know and deserve to know is what has happened, which is not partisan, which is not subject to debate, and then people can make their own guesses about where things are moving, where they will go and what they think should be done. And I think that’s fundamental. And we still live in this age of people throwing out words like fake news and alternate facts — and that’s just not OK. It’s never been OK.”

Here and Now, NPR

25 Open Datasets for Data Science Projects

Some of the best datasets for data science projects are those created for linear regression, predictive analysis, and simple classification tasks. This list will include the best resources from our past dataset articles tailored for said tasks. We’ll also highlight some of the best websites to search for open datasets on your own.

Lucas Scott, lionbridge.ai

London Busker Breaks it Down

Do you remember the last busker you saw on the tube? There’s a chance they remember you. If you had any kind of interaction with me, I would. If you smile, wave, cover your ears or sing along, I’ll record it.

One such notable interaction occurred at Leicester Square on a Saturday afternoon during a bog-standard rendition of Bruno Mars’s The Lazy Song. It has a catchy chorus which is preceded by the line, “I said it ’cause I can!”.

Just as I started the “…I can”, I noticed a huge group of school kids approach my busking pitch, staring with huge grins on their faces. As they approached, I watched in slow motion as their mouths opened. They all took deep breaths and in a moment my one-man performance was joined by a chorus of voices in perfect sync and harmony and we became an unexpected choir right there on the Underground.

Dan Hough / Londonist