How to actually measure/compute data quality

168 views Asked by At

I need to come up with data quality metrics for a project and how to measure them. I've been googling and reading and I understood that you can 'measure' the quality of data using the 6 dimensions (Accuracy, completeness, uniqueness, validity, timeliness and consistency).

For Accuracy, the 'ratio of data to errors' metric can be used. For Completeness the 'number of empty values' metric can be used and for Timeliness the 'data time to value' metric can be used (formulas of each below):

ratio to errors: tot nr of errors (errors = missing, incomplete or redundant entires) / tot nr of rows

nr of empty fields: tot nr of empty fields/NULLs in data set

data time to value = time data received - time data requested

For uniqueness, validity, and consistency however I did not find a formula. Can someone please help me? I've been googling forever, but I didn't manage to find concrete examples of how the above dimensions are actually measured. Thanks a lot! :-)

0

There are 0 answers