Blog

This week we kept working with normal distributions and what they actually look like in practice. We talked more about skewness and kurtosis and how they describe the shape of the data, especially how kurtosis can point out heavier tails or potential outliers. I feel pretty comfortable using z-scores now to standardize values and see how far things are from the mean. We also looked at the chi-squared distribution and the Jarque–Bera test as ways to check whether real data is close to normal instead of just assuming it is.

A p-value is the probability of getting results at least as extreme as the ones observed, assuming the null hypothesis is true.

APSTATSGUY explained in a very dramatic way what a p-value is. When the p-value is below .05 you question what is going on.

I will be using these ideas when working with real datasets to check assumptions, spot outliers, and decide whether a normal model actually makes sense before doing further analysis.

Week one was basically just a check-in and intro to how the class is going to run. We had a snow day Monday, and the rest of the week was pretty relaxed — mostly talking through the project format, picking ideas for what we might work on, and having some light discussion about what probability actually means and how we’ll use it. Nothing too heavy yet, just getting oriented and setting things up for the semester. I haven’t decided on a project yet but am considering something in healthcare. If I come up dry on that topic I can always refer back to my pitch prediction and work more in depth with that.