The idea of this project is to use data from google trends to predict a possible abnormality in the health status of the general population.
I assumed that people usually google their symptoms when they are experiencing a health issue. For example, a person that has a heavy cough may think that he has pneumonia and thus google ”pneumonia symptoms”.
As shown in Google Trends and as expected, people search for this term more in the winter than in the summer. That is of course because people experience more respiratory problems in the winter. Also, if we only consider the data before the COVID outbreak, we observe that roughly the same amount of people search for this term each year. In this project, I apply several Deep learning algorithms to predict the number of Google searches for the phrase ”pneumonia symptoms”.
The idea is that, if the actual searches outnumber the predic- tions, there might be something going wrong with the health of the general population, as more people have pneumonia-like symptoms than expected.
See a full analysis of the project in Time_Series_Forecasting.pdf