This is a repository about a Digital Epidemiology task regarding a machine learning model that try to predict the values of depression in USA (crude prevalence in region), from 2016 to 2018, according to different proxy data downloaded from google trends. In particular, as proxy data research trends of many words related to depression and the income values are considered.
In data2016.csv, data2017.csv, data2018.csv there are all the data. In DigitalEpidemiology.py there is the machine learning model written in Python.
The Continuous Integration with Travis was then performed in order to evaluate the aforementioned script on different Python versions, python 2.7 and python 3.6, and on different operating systems, linux and oxs.