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. On the other hand, social media also brings new forms of bias and shadings while it attracts people to circle information by its openness. Facing to newly forming technical lock-ins and the loss of local knowledge in agriculture in the era of plain digital transformation, the urge of re-establish a farmer-centric precision agriculture is evident. We want to know if social media like Twitter can help farmers to be heard. In this work, we chose several scenarios to collect Tweets, then we applied different natural language processing techniques to measure the informationality of Tweets in french as a complementary source for phytosanitary monitoring.