For decades, Americans spent more money at the grocery store than at eating and drinking establishments. It’s not like that anymore, Quartz reports.

Really fun. “Motion capture, procedural animation and dynamic simulations combine to create a milieu of iconic pop dance moves that become an explosion of colorful fur, feathers, particles and more.”

Even if there were a statistical model that predicted a mass shooter with 99 percent accuracy, that still leaves a lot of false positives. And when you’re dealing with individuals on a scale of millions, that’s a big deal. Brian Resnick and Javier Zarracina for Vox break down the simple math with a cartoon.

If you look at gun death rates for other western countries and adjust for population, the United States is a sore-thumb outlier. Kevin Quealy and Margot Sanger-Katz for the Upshot report.

Be sure to look at the headline for a few seconds (if you’re on a desktop). It changes to provide different baselines to compare the US rate against.

Voter turnout and political leanings for various demographic groups play an important role on the campaign trail. Candidates can’t go everywhere and talk to every single person, so they pick and choose. From the voter perspective, turnout feeds into an indicator for influence. In this interactive by Nate Cohn and Amanda Cox for the Upshot, Democrat percentage is plotted against turnout.

Each bubble represents a demographic — such as Asian women between 45 and 64 years old with college degrees living in California — and size represents number of votes in 2012 or 2004. See the big picture at first, and then use the dropdown menus to filter down to your group of interest.

Geographers Seth Spielman and Alex Singleton used something called “geodemographic classification” to classify small areas based on demographic averages.

[F]or example, we can identify places dominated by small apartments occupied by single city dwellers from those family residences and larger detached homes. The techniques are very popular in industry for customer segmentation – with logic following that our purchasing behaviour is influenced by where we live.

So it’s not just mapping race, age, or housing individually. Instead, the method provides much more detailed and descriptive clusters. Then, CartoDB recently made an interface to search and browse the data.

The Thinking Machine, by Martin Wattenberg and Marek Walczak, shows you the thought process of a computer trying to win at chess. There have been several iterations that date back to 2002, but the most recent iteration was built for modern browsers and you can play against the computer.

With an animated take on the commute map, Mark Evans shows where people commute to work.

The resulting animations are somewhat hypnotic (even my dog seemed to go into a trance watching them leading to minutes of human amusement) but also provide a visual way of quickly seeing the distribution of workers into a given city. The points are sized based on the number of commuters, so a large dot indicates a higher relative number of commuters moving from the same tract to the same tract. The dots are also color coded to see which counties are most represented in the commuter sample.

Just select a county to see. [Thanks, @Mikey_Two]

I hear there’s some show called “Game of Thrones” that’s kind of popular these days. Twitter visualized how every episode was discussed, counting the character connections, the emojis used, and the changes over time.

See how popular each character was, and the emojis used to described each character. In the visualization below, each circle represents a character with its size proportional to how often the character was mentioned in the Tweets and color representing affiliation of the character. The most used emojis for each character are displayed under the character name.

[Thanks, @kristw]

For most, crying isn’t an especially common occurrence over a long period of time, but when it happens, it’s often because something significant occurs in one’s life. Over the course of a couple of years, Robin Weis has 394 such occurrences. She knows this because she tracked when she cried and then later classified each event.
Read More

Terrapattern is a fun prototype that lets you search satellite imagery simply by clicking on a map. For example, you can click on a tennis court, and through machine learning, the application looks for similar areas.

Terrapattern uses a deep convolutional neural network (DCNN), based on the ResNet (“Residual Network”) architecture developed by Kaiming He et al. We trained a 34-layer DCNN using hundreds of thousands of satellite images labeled in OpenStreetMap, teaching the neural network to predict the category of a place from a satellite photo. In the process, our network learned which high-level visual features (and combinations of those features) are important for the classification of satellite imagery.

Give it a try.

Flight pattern maps are fun to look at and reveal the complexity of air transportation on a daily basis. But, there are other angles to look at this data from. Martin Grandjean used a force-directed graph to focus less on geography and more on volume and connections. Color represents continent, circles represent airports, and circle size represents number of routes.
Read More

Hilary and Roger touch on some interesting topics in the most recent Not So Standard Deviations, specifically on scalable and automated data analysis.

At the surface, it can seem like computers should be able to do the bulk of any analysis. Plug in the data, crunch the numbers in an algorithmic black box, and presto change-o you get a list of actionable insights. From that point-of-view, you should be able to build software that does almost everything for you. That’s almost never the case, and you realize it quickly once you dig into the data yourself.

It’s the same deal with visualization.

You see the end result, and it’s easy to imagine applying the same chart to another dataset. Geometry and color are easy to make with a couple lines of code. The chart should be generalizable, right? Sure, but the challenge is getting to that final point. There are various paths you can take when you start with a dataset — what it means, the questions you want to ask — along with various decisions along the way.

Automating the process. That’s the hard part.

Predictive policing seems to be playing a bigger role in court decisions these days. People charged with crimes can be given a risk score based on priors and their background, which represents a fuzzy likelihood that they commit a crime again. ProPublica investigates the reliability of these scores, using data from Broward County, Flordia, between 2013 and 2014
Read More