< CULTURE & MEDIA >

The other day I had to go see Dawn of the Planet of the Apes but the sun has diverted me from my goal. In the evening, I started in the (re) reading the original novel La Planète des Singes by Pierre Boulle published in 1963. I took the opportunity to make this poster or this cover: a drawing of all the sentences of the book.

Visualization is inspired by Sentence Drawings by Stefanie Posavec. Here we count the number of characters per sentence, not the number of words. TRinker wrote a post with the R program to achieve these own Sentence Drawings.  I used it. To have a “planet” effect, I use a polar plane rather than a Cartesian plane. The colors are essentially those of the first American publication, I did not like the colors of the French version.

According to my tests, the result is more attractive for relatively long texts. It is possible to play with several features: the length and the thickness of the lines; the color and the transparency of the lines; the angle of the segments, turn right or left, with 90° and not; and the coordinate plane, Cartesian or polar.

Okay this visualization is more artistic than functional. But that’s what I wanted to here :)

The drawing is done with R, I used the libraries ggplot2 and grid, and the TRinker’s program. I also wanted to take the opportunity to see the conditions necessary to import a .txt book in R.

Hi-Res here.

A small graph related to the previous poster on the book “Planet of the Apes” by Pierre Boulle. The book is the source of a media franchise consisting of films, books, comic books, video games, television series,… in a logic of spinoffs. I just took a look on the side of movies and series: date, type, ratings, reviews/critics, revenues (only from theaters)… No statistics, just a graph to keep a small trace of these notes.

The 1963 book has more reviews than movies and series :)

The 1968 movie (with Charlton Eston) is still very popular with the second highest rating. The 1970 serie was a commercial success, but less and less important over movies. It had a second life on television in the early 1980s. Spinoffs following (TV serie and Animated serie) are actually much less popular and lower-rated.

And 25 years later, if the Tim Burton movie is an economic success (audience and revenues), its rating is very low. But the new sage seems to live up to expectations: popularity + high ratings, users share and discuss their opinion in mass.

Always performed with R. The data come from Goodreads for the book and IMDB + Box Office Mojo for movies.

Hi-Res here.

Last month, the biggest sale of a comic took place: $ 3.2 million for the first appearance of Superman in Action Comics # 1… a perfect copy ! I take this opportunity to talk about current events and not only history of comics. 

The right side lists the 150 largest sales of comics based on the volumes: the issue, the cover date, the sale prices of the different versions (logarithmic scale), the number of sales, the biggest sales and sales amount. The data come from It’s All Just Comics.

The left side inserts the superheroes. I used Comic Vine API to extract characters by issue and to related issues. I distinguish the first and the second incarnations of superhero (Human Torch, Flash and Green Lantern). I selected some superhero for the following reasons: first appearance, first volume dedicated, as the lead story, or any other reason indicated. The size depends on the sales number in which the superhero appears.

I also put two colors to distinguish DC Comics & Marvel Comics who totally dominate those sales with Superman, Batman and Spider-Man. It is a little piece of the history of mainstream comics: DC Comics for the Golden Age, Marvel Comics for the Silver Age and the seconde incarnation of Flash between these two ages.

The set was made with R only, from extraction to graphics. I used the library rjson for Comic Vine API.

Hi-Res here.

The Signatures of Wikipedia. 287 versions of Wikipedia → 287 unique signatures. We cross the number of new articles published each month (vertical axis) with the numbers of contributors with 5 or more edits each month (horizontal axis). We introduce a mirror effect (dashed line): the growth of contributors (left) and the decline of contributors (right). To obtain this unique and fluid signature, we applied a polynomial smoothing to all data and we normalized all graphs. These data are from January, 2001 to June, 2014.

The graphs are performed in R.

Poster [1/4]Poster [2/4]Poster [3/4]Poster [4/4].

Hi-Res here.

The Signatures of Wikipedia. 287 versions of Wikipedia → 287 unique signatures. We cross the number of new articles published each month (vertical axis) with the numbers of contributors with 5 or more edits each month (horizontal axis). We introduce a mirror effect (dashed line): the growth of contributors (left) and the decline of contributors (right). To obtain this unique and fluid signature, we applied a polynomial smoothing to all data and we normalized all graphs. These data are from January, 2001 to June, 2014.

The graphs are performed in R.

Poster [1/4]Poster [2/4]Poster [3/4]Poster [4/4].

Hi-Res here.

The Signatures of Wikipedia. 287 versions of Wikipedia → 287 unique signatures. We cross the number of new articles published each month (vertical axis) with the numbers of contributors with 5 or more edits each month (horizontal axis). We introduce a mirror effect (dashed line): the growth of contributors (left) and the decline of contributors (right). To obtain this unique and fluid signature, we applied a polynomial smoothing to all data and we normalized all graphs. These data are from January, 2001 to June, 2014.

The graphs are performed in R.

Poster [1/4]Poster [2/4]Poster [3/4]Poster [4/4].

Hi-Res here.

The Signatures of Wikipedia. 287 versions of Wikipedia → 287 unique signatures. We cross the number of new articles published each month (vertical axis) with the numbers of contributors with 5 or more edits each month (horizontal axis). We introduce a mirror effect (dashed line): the growth of contributors (left) and the decline of contributors (right). To obtain this unique and fluid signature, we applied a polynomial smoothing to all data and we normalized all graphs. These data are from January, 2001 to June, 2014.

The graphs are performed in R.

Poster [1/4]Poster [2/4]Poster [3/4]Poster [4/4].

Hi-Res here.

I often use Wikipedia for my work, it’s a data mine. There are a few times, a series of articles noted the decline of Wikipedia. The article by Tom Simonite in the MIT Technology Review is very good and the academic paper by A. Halfaker, R. Stuart Geiger, J. Morgan and J. Riedl serves as a reference.

But I found there was something missing: the crossing between the number of articles and the number of contributors, two major characteristics highlighted in the two previous references. So I made this little graph to view the importance of the current problem. Wikipedia is a great project but it is not immortal. It is necessary to resolve this issue.

The graph is performed in R. The data come from Wikipedia. A brief methodological guide is located here, and an extension there.

Hi-Res here.

The brief methodological guide of the graph The Endless Decline of Wikipedia.

Hi-Res here.

A methodological extension to the main graph The Endless Decline of Wikipedia. The idea is to apply this method to all versions of Wikipedia and get an unique and fluid signature.

Hi-Res here.