Epitweetr is an R package including an interactive tool to help automated early detection of public health threats using Twitter data. Francisco Orchard, head of Epiconcept’s Data Sciences business, and Esther Kissling, deputy director of the Department of Epidemiology, contributed to the project. A version 2 of the tool is currently under development, with a team completed by Lore Merdrignac, epidemiologist and biostatistician. It will propose in particular :
– An adaptation to the evolutions of the social network: in 2019, Twitter launched version 2.0 of the Twitter search API, which will eventually replace the current version 1.1. Epitweetr will now include an additional option to use this new version of the API.
– A revision of the geolocation algorithm, to provide better accuracy.
– The addition of a machine learning algorithm for automatic signal classification.
– New features: progress bar in the dashboard, improvements to email alerts, etc.
– Search capabilities on all collected tweets
– Improved overall performance of the tool.