Visualizing Rank Time Series of Wikipedia Top-Viewed Pages

Abstract– Large rank time series datasets have alwaysposed a challenge for data analysts. In addition to the time-varying property, each itemat every time point can be a complex object that has multivariate properties, relation properties, or both. In this article, we study rank time series to explore their evolving patterns and the relations among the top items. The online encyclopedia Wikipedia has a daily page-view throughput of more than 600 million across all available languages (see stats.grok.se). It has become a knowledge-gathering platform where users can learn and contribute their own knowledge. Due to this sustained large-scale knowledge accumulation, Wikipedia has also become a huge and growing knowledge warehouse. It would take more than a human lifetime to digest all the knowledge collected in Wikipedia. But to understand only current news and events,
we could just read through the top-viewed pages. According to the 80/20 principle, the ranking data of Wikipedia’s top page-view statistics (Wikipedia page rank, for short) reflcts users’ major interests in Wikipedia, and its top queries further indicate ongoing affairs. Therefore, the time series of Wikipedia page rank (Wikipedia page ranking trends) indicate how users’ interests evolve over time