Below are projects created as part of an assignment for UC Berkeley's Professional Certificate in ML & AI. You can find them in their own folder.
A business is delivering coupons to drivers for food-related establishments in their proximity. Our goal for this study is to determine the factors that affect the coupon acceptance rate. We use a number of different techniques, including exploratory data analysis and statistical tests, in our analysis.
The used car market is one of the largest in the world. Used cars are in high demand because they are typically sold at discounted prices while remaining usable.
In this exercise, a used car dealership wishes to develop a model using the CRISP-DM methodology to analyze used car sales. We develop and analyze a model to provide insights on how it could make a business more profitable.
A financial institution based in Portugal has collected data for over 41,000 telemarketing calls made over a five-year period. It wishes to use this data to predict whether or not a sales lead may become a client in the future. Given the opportunity cost of missing out on a potential customer, the ultimate goal is to minimize the number of false negatives.
Because each lead either is or is not a customer, this is a binary classification problem. We implement several different models (such as k-nearest neighbors, logistic regression and support vector machines) to find the one with the highest predictive performance, and we also investigate methods to further optimize the models.
Simply upload the Jupyter Notebooks and the accompanying data files to your Google Drive, and then open the Notebook in Google Colab. You will need to update the path to the data file or use the same directory structure. Please see the Google Colab tutorial for more information.