In this chapter, we looked into unsupervised machine learning techniques with Julia. We focused on clustering, one of the most widely used applications of unsupervised learning. Starting with a dataset about businesses registered in San Francisco, we performed complex—but not complicated, thanks to Query—data cleansing. In the process, we also learned about metaprogramming, a very powerful coding technique and one of Julia's most powerful and defining features.
Once our data was in top shape and after mastering the basics of clustering theory, using the k-means algorithm, we got down to business. We performed clustering to identify the areas with the highest density of companies to help our imaginary customer, ACME Recruiting, to target the best areas for advertising. After identifying the parts of the city that would give ACME the best reach, we performed data analysis to get the top domains of activity required by our customer so they could build a database of relevant candidates...