Data pipelines
Even with the most basic of analytics, we always require some data. In fact, finding the right data is probably among the hardest problems to solve in data science (but that's a whole topic for another book!). We have already seen in the last chapter that the way in which we obtain our data can be as simple or complicated as is needed. In practice, we can break this decision down into two distinct areas: ad hoc and scheduled.
- Ad hoc data acquisition: is the most common method during prototyping and small scale analytics as it usually doesn't require any additional software to implement. The user acquires some data and simply downloads it from source as and when required. This method is often a matter of clicking on a web link and storing the data somewhere convenient, although the data may still need to be versioned and secure.
- Scheduled data acquisition: is used in more controlled environments for large scale and production analytics; there is also an excellent...
