Team member: We're going to need some starter data to normalize our standards. Where do we get that information? You: Our company has a lot of its own data we can use, and we can expand those numbers using the SelectU.S. Census BureauGapminderFacebook GraphHealthdata.govItem 1 site. Some of these metrics, though, will have to be normalized as we collect data ourselves.

Team member: It's such a huge variety of data types. How are we supposed to work with all of this? You: We have ways of handling SelectstructuredhistoricalunstructuredmedicalItem 1 data that accounts for the variety and even benefits from it. It's challenging at the beginning, especially in the design phase, but there are rich insights available when it's done well.

Team member: I assume we'll be storing this data in the cloud? Using IoT devices like these fitness trackers means the data will be traveling over the Internet anyway. Might as well drop it into a cloud-based database. You: Right. We'll need to ensure the storage space is SelectGDPRHIPAAPCIPATRIOT ActItem 1 compliant, but that shouldn't be hard to find these days.

The choice of a relational database management system versus a NoSQL solution depends on the problem that needs to be addressed. Often, the data structures used by NoSQL databases are more flexible than relational database tables and, in many cases, they can provide improved access speed and redundancy.