In most cases with Big Data projects you have to ingest 6-10x more data than has any immediate value just to start and glean value. This creates a large volume of data with no apparent value, and sometimes it is a long time before there is any value if ever. This creates a large expense and a lot of issues for entry, which has derailed  a number of big data projects. So be prepared going in that unless you really know your data you will need to ingest and explore to find the value. When you do the numbers for your project keep this in mind.