Unveiling Insights Exploratory Data Analysis Essentials

Assignment Task

1. Exploratory data analysis: check missingness, possible errors in the data, explore relationships, look for signs of confounding and biologically plausible interactions

2. Build a causal diagram that connects the different variables to each other based on your subject knowledge

3. Build your regression model:

a. Build univariable regression models for each predictor and select those with a p< 0>

b. Build your multivariable model

i. Backward elimination: add all variables selected at step 3.a together in a multiple model, then remove one by one starting from the highest non-significant value. Pay attention to the change in coefficient if your question is about estimating the effect of a particular predictor e.g. treatment on disease. Continue until all variables left in the model are all significant or an important confounder. At the end, try putting back each variable you have removed one at a time to see if they still are not significant

ii. Forward selection: add the variables one by one. If it is significant and/or changes the coefficient of your predictor of interest keep it, otherwise discard it. At the end, try to put back each discarded variable you have removed one at a time to see if they are still not significant

This Data Analysis