How much information is in New York Times articles ?
|Jan Chołoniewski , Robert M. Paluch , Krzysztof Suchecki , Janusz A. Hołyst|
Warsaw University of Technology, Faculty of Physics, Koszykowa 75, Warszawa 00-662, Poland
We present results of an extensive analysis of data set containing 6.5 million articles published in New York Times. Using a cosine similarity measure describing co-occurrences among words in the articles we have constructed a directed network dependent on time-dependent historical framework of the considered data set as well as on individual attractiveness of published papers. We have estimated a total volume of all trees in the constructed network that were started on a given day. A large volume means that an article or group of articles published on this day launched long-living information threads. The maximum volume has been found for days related to the beginning of Great Depression in October 1929 (Wall Street Crash on Black Tuesday). A distribution of informational explosiveness of individual articles has been also studied. The explosiveness has been defined as a ratio of number of articles that follow a given articles to a number of articles that were followed by this article. The distribution is close to a power law with a characteristic exponent 3.5.
Presentation: Invited oral at 8 Ogólnopolskie Sympozjum "Fizyka w Ekonomii i Naukach Społecznych", by Janusz A. Hołyst
See On-line Journal of 8 Ogólnopolskie Sympozjum "Fizyka w Ekonomii i Naukach Społecznych"
Submitted: 2015-09-09 00:17 Revised: 2015-09-09 00:20