In this project, we will work with real-life data provided to us by Udacity's Bertelsmann partners AZ Direct and Arvato Finance Solution. The data here concerns a company that performs mail-order sales in Germany. Their main question of interest is to identify facets of the population that are most likely to be purchasers of their products for a mailout campaign. Our job as a data scientist will be to use unsupervised learning techniques to organize the general population into clusters, then use those clusters to see which of them comprise the main user base for the company. Prior to applying the machine learning methods, we will also need to assess and clean the data in order to convert the data into a usable form.
- NumPy
- pandas
- Sklearn / scikit-learn
- Matplotlib (for data visualization)
- Seaborn (for data visualization)
The unsupervised learning branch of machine learning is key in the organization of large and complex datasets. While unsupervised learning lies in contrast to supervised learning in the fact that unsupervised learning lacks objective output classes or values, it can still be important in converting the data into a form that can be used in a supervised learning task. Dimensionality reduction techniques can help surface the main signals and associations in your data, providing supervised learning techniques a more focused set of features upon which to apply their work. Clustering techniques are useful for understanding how the data points themselves are organized. These clusters might themselves be a useful feature in a directed supervised learning task. This project will give us hands-on experience with a real-life task that makes use of these techniques, focusing on the unsupervised work that goes into understanding a dataset.