Separating Treasure from Trash: Quantifying Data Waste in App Reviews

User generated app reviews provide valuable information to multiple stakeholders, however, the exploding number of reviews creates practical business and environmental problems. We propose a new approach for making reviews more environmentally sustainable - reducing waste at source. Many app reviews are 'trash', containing little or no potential information value. Alternatively, some reviews are 'treasures' providing actionable information. Using these definitions, we develop an automated method to distinguish between trash and treasure app reviews in the Google Play entertainment app category. We find 15% of app reviews are pure trash, while only 26% represent true treasure. Reducing trash reviews at source across all app categories in four major app stores could result in reductions of CO2 emissions ranging from 128 kg to 608 kg depending on the waste-reduction approach adopted. We offer suggestions for eliminating waste at source, including changes to the review interface and acceptance process.