The vocabulary used in media (e.g. news) and social media (e.g. Twitter) on a disease changes according to the period. In this context, text-mining and terminology extraction tasks can be used to analyse epidemic periods of diseases. Moreover, we have to take into account this knowledge in order to improve event-based surveillance (EBS) systems. Text-mining and machine learning approaches can be integrated in different steps of EBS systems for disease-based and symptom-based surveillance: data acquisition, information retrieval (i.e. identification of relevant documents), information extraction (i.e. extraction of symptoms, locations, dates, diseases, hosts, etc.), and visualisation. This work highlights the use of text-mining approaches related to COVID- 19 (i) for surveillance systems (i.e. web crawling and information extraction tasks) and (ii) for spatio-temporal analysis of tweets.