Getting and cleaning data

If you work in an organization, you probably already have access to some datasets used by the IT systems. However, acquiring them and formatting them is not all that simple. Some clear requirements need to be specified regarding the valid data range – that is, the list of characteristics available and/or needed. Getting a usable dataset can require several iterations with the data provider. At each iteration, some data quality checks need to be performed. These checks, of course, depend on the problem to be solved, but some of them are mandatory. For example, you always need to ask whether it's good quality data and whether you need more.