Beautiful in English


Project description

February & March 2017


data sketch|es collaboration with Google


Excel, R, D3.js


Google translate data

While on vacation Shirley Wu, my partner on data sketches, and me got an email from Alberto Cairo. If we were perhaps interested in doing a “data sketches” style visualization project together with Google News Lab. Well yes, of course we were! It was the project that finally convinced me to go into freelancing :)

We were completely free to come up with our own angle, as long as it was based on Google data. During November and December we brainstormed several ideas and finally Simon Rogers from Google News Lab chose the topic of Culture. As a non-native English speaker who visits Google Translate at least once a day, I was very interested to find out what other languages translate into English and if there are any cultural differences (spoiler: there are). During January I worked together with Google News Lab and the Google Translate team to get a dump of the single word translations from 10 chosen languages into English. It was still a lot of work to clean up these words, since I was only interested in the nouns and adjectives.

I focused on creating a page that would have a “top-down” approach. Starting with the most common noun or adjective at the top, which just happened to be beautiful, I was very very happy with that, and going down into more detail the further down you scroll. From the most translated word per language, to the top 10 per language and finally to the similarities between languages.

For each of these three sections I created a visualization that was heavily build up of words itself, fitting the topic of translations. The first two visualizations actually consist of nothing else then words :)

This was also one of my first projects where I created it to work perfectly fine on mobile. Several layout changes happen to the visuals, as compared to the desktop version, to make optimum use of the rectangular screens of mobiles.

It was a lot of fun to work on this project and dig into a dataset that was, for once, not available online, but specifically pulled out of Google’s systems :D

You can a lot more about the creation process here


The most Translated noun or adjective per Language

The first visual highlights the most often translated noun or adjective for 10 of the most popular languages on Google. Stringing all languages together is a string of the overall top 100 most often translated noun+adjectives. You can hover over any of the languages to see the Google Trend for the English word and some interesting related queries and topics connected to the word.

The wordsnake visual of the most common translated words

When moving to smaller screens, the layout of the languages changes to make optimum use of the available screen width without getting smaller in size

The top 10 per Language

Next, we dive into the top 10 words per language. On top there is a large “tree-ring” like circle. The English translation is the bigger black word in the center and the smaller grey words to either side give the original word. You can switch between any of the tiny language “flowers” below to see it rotating into place in the bigger circle.

The top 10 per language tree ring visual

Similarities between Languages

The final visual focuses on the similarity between languages. Any word that two languages share between their top 10 is visualized by a connecting line. Some languages, such as Spanish and Portuguese, have 4 words in common. Whereas Japanese has virtually nothing in common with any other language.

You can click on each language circle to make it move towards the center. This will reveal the actual words by which it is connected to all the other languages.

Switching between languages in the similarity network

Like the first “wordsnake” visual, this network also changes shape when moving to smaller screens. On desktop it is a perfect circle, whereas on mobile it forms more of a rectangle, taking up as much space as available to keep any word overlap as small as possible.

Language similarity network - mobile versus desktop