The life cycle of ideas is a data-analysis and visualization we published for the May Issue of Popular Science.
As Katie Peek wrote in the introductory text:
“Every scientific idea has its day. Theories are born and experiments are designed; results are put to the test, then disproved or accepted as canon. As scientists discuss an idea, they cite the paper that proposed it in their own work. Then, as the conversation moves on, references to the paper drop off. The rise and fall of citations serves to measure the lifespan of a paper’s underlying ideas. Popular Science visualized that pattern across disciplines. Generally, citations peak more quickly today than they did 50 years ago. According to Jevin West, an information scientist at the University of Washington, that trend could be because there are more scientists tackling problems, or because technology has connected them better, accelerating the conversation.”
A little introduction on the data and on our methods of analysis follows, to pave the basis to explain our visual design choices.
(high resolution image here)
The visualization represents scientific papers and their peaks of citations over years.
Using the Web of Science database, we considered the 20 largest science fields—those with the most articles each year, and within those fields, we calculated how long it took an average of the 50 most-referenced papers published each year to reach peak citations. In order to reduce possible biases due to the low number of citations for more recent publications, the dataset contains papers published until 1998. For these papers we had citation data for every year until 2008 (about 12 million records)
After the selection of the disciplines, we first calculated for each paper in each discipline the number of years that passed from its publication date to the year where the number of citations of that paper has its maximum value: we considered this measure as a proxy for the time it took every paper to reach “success”.
Where we had the same maximum number of citations in two different years, we picked the oldest one (i.e. the “earliest success”). Having this data, we could calculate for each discipline and year the average number of years it took for the top 50 papers published in that year to reach the maximum number of citations.
We represented this as a spark-line following the circumference of the field. Finally, we picked the top 20 papers in each field based on their cumulative number of citations, and we represented those as “notes”, throughout the circumference of the field they belong to.
Each note is positioned in the circle according to the year of publication of the paper. The length of the note corresponds to the number of years it took that paper to reach the maximum number of citations. In addition, for each paper we calculated the cumulative number of citations it received since it was published and represented this as well as a small bar springing out from the base of the note towards the center of the circumference. The number of authors for each paper and recurrent authors are visualized as well (as small dots and arches connecting them) .
We precisely referred to the musical notation panorama, as you can see in most of these references we have been looking at before starting, but then we also combined the main design inspiration with other “visual worlds” to compose these articulated “musical trees” of scientific disciplines, respecting the rich data analysis we performed but still evocating familiar visual panoramas.
Giorgia Lupi, Simone Quadri, Gabriele Rossi, Glauco Mantegari, Pietro Guinea Montalvo, Davide Ciuffi