Indirectly, yes. The Coursera course teaches K-means clustering as an example of an unsupervised learning algorithm, along with a technique to work with larger

Indirectly, yes. The Coursera course teaches K-means clustering as an example of an unsupervised learning algorithm, along with a technique to work with larger data sets in a more manageable way (e.g. PCA to reduce dimensionality). With the basic building blocks from the ML course you have a much better understanding of what Google and Stanford did. Andrew Ng is also the course instructor for the Coursera ML course[1]. I am presuming that supplemental readings and Youtube videos by Prof. Ng (e.g. http://www.youtube.com/watch?v=ZmNOAtZIgIk http://www.youtube.com/watch?v=ZmNOAtZIgIk) allows a sufficiently motivated individual to at least partially replicate the results in the OP. [1] Andrew Ng is also a co-founder of Coursera.