Week 57

1/08 - 01/12: ## done and to do * SFI Application * CHOP Models * train on precovid, test on covid * new stratification based on number of visits, race, gender * spent full day writing code to merge precovid and covid dataframes (now elina should be able to use this file too) * summary results to answer questions: how many people convert to suicide attempt after having baseline of no suicide attempt, and which zipcodes have highest suicide attempt, demographic groups most likely to attempt suicide, anxiety and depression levels pre- and postcovid * figure out how to optimize secondlevel modeling – currently routine for identifying top features produces variable results