Dataflow Modeling

1. Refresh Periodicity
How often is this data file refreshed? On what basis do I get it? Is it going to be monthly on the first Tuesday? Is it going to be every other Wednesday at 1pm? Can I get it on demand?

2. Data Periodicity
The data contents of the file. Are they monthly, weekly, daily data?

3. Fallback & Dependencies
Is there an alternative source for this data? If this file does not arrive, what processes are affected? If this data needs to be resubmitted what is the guideline? How often can it be resubmitted? Does the target need to be reset?

4. Data Owner
Who is responsible for this data? Who understands what it means?

5. Volume
How big is the file? What is a typical size for the file?

6. Data Slice
What is the cut of this data? Is it in an archiveable form? Why is the data cut in this way? Is it along the dimensionality of the data? Is it a complete dataset? How does the physical cut of the data related to the logical cut of the data? Is it consumed in the same grain as it is produced?

7. Data Target
What is the consumer of this data? Does the target need to be reset or is this incremental data?

8. Format
Is this CSV? Fixed length or delimited fields? How many lines of header? Field names on top? Record length? Delimiter? What is the file spec & naming convention for this file?

9. Notification
How do I know about the existence of the file? What will I know about the file before I get it? What can I know about the file before I read it?

10. Data Increment
Is data restated? Should it overwrite the target data or be added incrementally. Are these balances or transactions?