Risk of a large sample size Typically we don’t worry about sample size being too large, because that means we have more power, we can detect smaller effect with more significance. It’s not a problem because of technical reasons but human motivation and interpretations. Warning: Huge Samples Can Make the Insignificant…Significant When sample size is too large, unless the effect is truly not there, unless the coefficient is truly zero, it doesn’t matter how small the beta is, we will find it significant.

Introduction Unfair Competitive Advantage When watching competitive swimming, it seems like the swimmers in the middle lanes always win the race. Although this is partly by design, as the fastest qualifiers are placed in the innermost lanes, there is great debate amongst the swimming world whether there is additional advantage unintentionally given to these swimmers. In this project, we hope to address whether lane placement leads to an unfair competitive advantage.

Summary Background Star Digital, a large multichannel video service provider, spends a large portion of their budget on advertising. As the technological environment changed, so did Star Digital’s advertising strategy as they began to invest more heavily in online advertising such as banner ads. In order to get the most out of their budget, they actively evaluate the return on investment (RoI) of each ad medium. In evaluating the effectiveness of an online advertising campaign, Star Digital has designed a controlled experiment.

Honestly, I don’t care what we do tonight. We could go out and get drinks, or grab some beer and stay in. We could check out that new Japanese place we’ve been meaning to try, or just swing by that Mexican place by the river. We could make Chinese as always. We could get pizza on the patio. We could pound down a two-gallon bucket of tzatziki sauce in the back of the car.

MSE Sensitive to outliers Has the same units as the response variable. Lower values of MSE indicate better fit. Actually, it’s hard to realize if our model is good or not by looking at the absolute values of MSE or MSE. We would probably want to measure how much our model is better than the constant baseline. Disadvantage of MSE If we make a single very bad prediction, taking the square will make the error even worse and it may skew the metric towards overestimating the model’s badness.

Image classification exercise I’m not gonna lie. I’m more of a dog person. Better know how to tell the difference. 😂 Here we go. Introduction This is an Image classification exercise. I will play with an expired Kaggle competition. The code is available in a jupyter notebook here. You will need to download the data from the Kaggle competition. https://www.kaggle.com/c/dogs-vs-cats/. The dataset contains 25,000 images of dogs and cats (12,500 from each class).