Continuing the neural network explorations, Shan Carter and team of Google Brain and Cloud, look at how a network deals with handwriting by placing them in the same space.

The black box reputation of machine learning models is well deserved, but we believe part of that reputation has been born from the programming context into which they have been locked into. The experience of having an easily inspectable model available in the same programming context as the interactive visualization environment (here, javascript) proved to be very productive for prototyping and exploring new ideas for this post.

Side note: Been seeing a lot of Google experiments the past couple of weeks. I like it. Is it because it’s December, or are they just feeling more experimental these days?

Geographers Alasdair Rae and Garrett Nelson used commuting data from the American Community Survey to identify “megaregions” in the United States:

The emergence in the United States of large-scale “megaregions” centered on major metropolitan areas is a phenomenon often taken for granted in both scholarly studies and popular accounts of contemporary economic geography. This paper uses a data set of more than 4,000,000 commuter flows as the basis for an empirical approach to the identification of such megaregions.

BuzzFeed used interpretive dance to describe the average age of the milestones in our lives, from birth, losing the first tooth, marriage, and death. The data points serve more as background, as a way to provide a timeline of events, and the dancing is the primary focus.

[arve url=”https://www.youtube.com/watch?v=9bvOHZ0NWig”/]

I found myself drawn to the comments on YouTube. Typically a cesspool of idiocy and more idiocy, the comment section in this case might be a good representation for how a (younger) general audience interprets averages. All of the top comments are basically, “I guess I’m not average” and “There’s no way that’s the average. [Insert comparison to self.]”

This of course is because averages are just that. They’re the sum of all individuals divided by the total population, and average values represent one aspect of a range or distribution of things.

So in the case of these average ages, most people either fall below or above instead of right in the middle.

But I digress.

[arve url=”https://www.youtube.com/watch?v=kIID5FDi2JQ”]

We’ve seen many one-off projects that show the distortions you get when you project a map. There’s just no avoiding them, when you convert a 3-D object onto a two-dimensional plane. Vox demonstrates and explains with an inflatable globe.

A few years back, Google released a time-lapse feature in Google Earth that let you see change through satellite imagery. They updated the feature last week. It’s more detailed and higher resolution than the first version, based on the pixels from about five million images.

We took the best of all those pixels to create 33 images of the entire planet, one for each year. We then encoded these new 3.95 terapixel global images into just over 25,000,000 overlapping multi-resolution video tiles, made interactively explorable by Carnegie Mellon CREATE Lab’s Time Machine library, a technology for creating and viewing zoomable and pannable timelapses over space and time.

Pretty cool to see my own neighborhood develop into what it is today.

Give it a go.

Update: See also The New York Times’ take on some of the water bodies around the world, using the same data.

Food trends come and go. Some stay longer than expected, and others come back a certain time every year. With their new project, The Rhythm of Food, Google News Lab and Truth & Beauty explore these patterns through twelve years of search trends.
Read More

I can always find time to enjoy me some minimal maps. Tim Meko for The Washington Post visualized American infrastructure in a series of six maps, from the electric grid to bridges and railroads. The above is pipelines.

Datalegreya is a new typeface by Fig that lets you intertwine a graph into the words.

Datalegreya can be used in all contexts where small space is available to synthetically display graphical data: connected objects, embedded displays, annual reports, weather report, stock prices, etc. It doesn’t need any specialized software: users just have to install it in the operating system and launch any software able to display OpenType fonts, such as Microsoft Word, Apple TextEdit, Adobe Suite etc. Standards compliance makes it equally available on the web or software embedded.

I rarely find myself searching for more space on the computer screen, and on the phone, I’d rather just scroll further down. Maybe useful on a smartwatch? Dashboard?

There was a lot of talk about deporting millions of illegal immigrants immediately, but as The New York Times shows, the actual number that could be deported is much less. Haeyoun Park and Troy Griggs use a clustered, force-directed graph to show the pool of 11 million immigrants and then filter down as you scroll. The transitions are key.

Here’s a short illustrated animation on introverts and extroverts by Julia Rodrigues.

[arve url=”https://vimeo.com/164177132″]

This is basically me. I enjoy the company of friends and family, but I do feel like I need to recharge my battery at times with some alone-time.

Dan Keating and Laris Karklis for The Washington Post map the change in diversity since 2000. The color scale, shown in the top right, represents two things: level of diversity and change in diversity. I’m not so sure the dual scale is interesting as a whole, as my brain just wants to split out each category individually or see each one separately. But keep scrolling and you can get that separation, which is a lot more visually helpful.

See also: percentage of white people, majority minorities, and predominant race.

Different species of birds make different sounds. However, the sounds are so quick and compressed that it can be tough to pick out what is what. So Kyle McDonald, Manny Tan, and Yotam Mann created a “fingerprint” for each bird song and used machine learning to classify. Through the visual browser, you can play sounds and search for bird types. Similar sounds are closer to each other.

In the latest addition to the put-big-numbers-in-context genre, here’s the history timeline of our planet in the context of 100 yards.

[arve url=”https://www.youtube.com/watch?v=M8V_glRW1hA”]

What started as a personal project and then turned into a book, Dear Data was a collaboration between pen pals through data. Now Giorgia Lupi and Stefanie Posavec’s work is headed to the Museum of Modern Art’s permanent collection.

[O]ur 104 original postcards and the many sketchbooks we filled with intermediate data drawings every week have found the best possible home for years to come. They will live in the archives and catalogue of one of the world’s most prestigious institution (well, we think so, at least!), humbled by being in the presence of the dazzling company of the great masters of art of the past two centuries.

This is amazing. Congratulations to Giorgia and Stefanie.