Informativeness in Twitter Textual Contents for Farmer-centric Plant Health Monitoring

Abstract

Data mining in social media has been widely applied in different domains for monitoring and measuring social phenomena, such as opinion analysis towards popular events, sentiment analysis of a population, detecting early side effects of drugs, and earthquake detection. Social media attracts people to share information in open environments. Facing the newly forming technical lock-ins and the loss of local knowledge in agriculture in the era of digital transformation, the urge to re-establish a farmer-centric precision agriculture is urgent. The question is whether social media like Twitter can help farmers to share their observations towards the constitution of agricultural knowledge and monitoring tools. In this work, we develop several scenarios to collect tweets, then we applied different natural language processing techniques to measure their informativeness as a source for phytosanitary monitoring.