This paper examines the causes of failures in automated news production. Although technical factors, such as bad data quality, must be considered, the human factor remains essential within its social dimension. The proposed approach is both theoretical and empirical. It is highlighted with the results of a one-year case study. The project involved developing an automated news application about air quality within a Belgian newsroom to provide audiences with real-time information and raw material for investigative purposes.
This abstract was presented through a poster at the Computation + Journalism Symposium, held at the University of Miami in February 2019
(Click on the image to enlarge it).
