Understanding the Gap between Forecast Accuracy and User Perceptions: Case Study of Japanese Weather Apps

There are numerous weather forecast apps available for smartphones in today’s market. While users will likely consider various features when choosing an app, forecast accuracy is considered one of the most critical factors. This paper presents a case study comparing the year-round forecast accuracy and user perceptions of two major weather forecasting apps in Japan: Yahoo! Weather and Weathernews. Our findings indicate that while the actual accuracy of the two apps is comparable, there is a notable difference in user perceptions of accuracy. Subsequent semi-structured interviews with 12 weather forecast app users revealed that there is a gap between users’ understanding of the forecast and the forecast provider’s intent, which may affect user impressions of accuracy. We categorized the instances where users felt the forecast was inaccurate into six patterns, that include emerging issues resulting from the recent diversification of functions within apps and the increasing opportunities to check weather forecasts from multiple types of devices (e.g., smartphones and smartwatches). We then discussed possible improvements for the user interface and presentation of forecast apps to address these gaps.