Data In, Fact Out: Automated Monitoring of Facts by FactWatcher

Towards computational journalism, we present FactWatcher, a system that helps journalists identify data-backed, attention-seizing facts which serve as leads to news stories. FactWatcher discovers three types of facts, including situational facts, one-of-the-few facts, and prominent streaks, through a unified suite of data model, algorithm framework, and fact ranking measure. Given an appendonly database, upon the arrival of a new tuple, FactWatcher monitors if the tuple triggers any new facts. Its algorithms efficiently search for facts without exhaustively testing all possible ones. Furthermore, FactWatcher provides multiple features in striving for an end-to-end system, including fact ranking, fact-to-statement translation and keyword-based fact search. 1. MOTIVATION Computational journalism [1, 2] is a young field that assists journalism using computing. One of its objectives is to find news leads backed up by hard, factual data. In the last several years, our research in this thrust has been focused on automatic and algorithmic fact finding by database and data mining techniques [3, 5, 4, 6]. Specifically, we studied how to monitor three types of facts that can be expressed as the following factual statements: Situational fact [4] “The social world’s most viral photo ever generated 3.5 million likes, 170,000 comments and 460,000 shares by Wednesday afternoon.” (http://www.cnbc.com/id/49728455) A situational fact is about a contextual skyline object within a certain context (e.g., all photos posted to Facebook) with regard to several measures (e.g., number of “likes”, number of “comments”, and number of “shares”), i.e., the object is not dominated by any object in the context when they are compared by the measures. One-of-the-few [5] “Victor Oladipo scored 30 points and handed out 14 assists ... only three other rookies have recorded at least 30 points and 14 assists in a game ...” (http://espn.go.com/espn/elias? date=20140222) This statement is about a one-of-the-four object, which is only dominated by at most three other objects. Prominent streak [3, 6] “This month the Chinese capital has experienced 10 days with a maximum temperature in around 35 degrees Celsius—the most for the month of July in a decade.” (http:// www.chinadaily.com.cn/china/2010-07/27/content 11055675.htm) A prominent streak is a long consecutive subsequence (e.g., 10 days of This work is licensed under the Creative Commons AttributionNonCommercial-NoDerivs 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/3.0/. Obtain permission prior to any use beyond those covered by the license. Contact copyright holder by emailing [email protected]. Articles from this volume were invited to present their results at the 40th International Conference on Very Large Data Bases, September 1st 5th 2014, Hangzhou, China. Proceedings of the VLDB Endowment, Vol. 7, No. 13 Copyright 2014 VLDB Endowment 2150-8097/14/08. Situational Facts (<*,*,DAL>,{pts,reb}) One-of-the-Few (, {pts, reb}) Prominent Streaks (, {pts}) Facts id player team ... pts reb ... t1 Larmar Clippers ... 12 9 ... t2 Larmar Clippers ... 8 11 ... ... ... ... ... ... ... ... t7 House Heat ... 8 6 ... t8 Larmar Clippers ... 10 11 ... Algorithms