Automating, Operationalizing and Productizing Journalistic Article Analysis

Public Good Software’s products match journalistic articles and other narrative content to relevant charitable causes and nonprofit organizations so that readers can take action on the issues raised by the articles’ publishers. Previously an expensive and laborintensive process, application of machine learning and other automated textual analyses now allow us to scale this matching process to the volume of content produced daily by multiple large national media outlets. This paper describes the development of a layered system of tactics working across a general news model that minimizes the need for human curation while maintaining the particular focus of concern for each individual publication. We present a number of general strategies for categorizing heterogenous texts, and suggest editorial and operational tactics for publishers to make their publications and individual content items more efficiently analyzed by automated systems. 1.Introduction: The Basic Problems In 2013, Public Good Software (PGS) introduced a product designed to direct readers concerned with the causes they read about in journalism and other media with the nonprofit organizations working on those causes. For example, an article about a shooting in Chicago would link to local anti-violence organizations so a reader could donate, volunteer, or just spread the word. This product, called “Public Good for Media” (PGM) initially debuted as a Javascript-powered “Take Action Button” that publications could insert into the HTML source of any page (see Figure 1). This first iteration of the product required explicit configuration of each insertion by a publication’s editorial staff with the assistance of PGS experts in curating the PGS marketplace’s collection of organizations and causes. Explicit configuration quickly proved untenable for both the publications and PGS in terms of staff time and expense, publishing workflow modifications and lack of growth in traffic driven through the PGM program. Figure 1: The Take Action Button embedded in an article To address this, PGS created tools to automate the process so that the PGM product could be inserted into a site-wide template and relieve the media outlets’ personnel of the burden of remembering how and when to include the PGM program as part of the publication process. Publishers also requested that PGS provide methods of automatically assigning the button’s destination based on the content of each individual story, as well as the suppression of the button’s display when there was no sensible destination, or when a story was not in any way cause-related (entertainment and sports, for example). Finally, while directing readers to nonprofits with a national area of effect was often sufficient, many publications either had more specifically located audiences, or wanted to highlight organizations working in the particular geographic area mentioned in the article. Bloomberg Data for Good Exchange Conference. 24-Sep-2017, New York City, NY, USA.