Summary

In this chapter, we covered topics related to data comparisons. We started by discussing the idea of a data fingerprint. In order to illustrate a data fingerprint, we introduced a hash function that can take any string and convert it to a number in a fixed range. This kind of hash function is useful because it enables us to ensure that data has the right identity – just like a fingerprint in real life. After introducing hash functions, we talked about the need for data signatures. A data signature or analytic signature is useful because it enables us to see whether two datasets are approximate matches – fingerprints require exact matches. We illustrated the use of analytic signatures with image data. We concluded the chapter by covering latent variable models. The latent variable model we discussed was factor analysis. We discussed an application of factor analysis that uses psychological survey data to determine differences between people's personalities. In the next chapter, we...