| Data Profiling provides a quick and easy way to | | | | business which may not otherwise be addressed |
| get a great understanding of your data with only | | | | by the original project |
| a handful of simple checks and techniques. And | | | | - Next steps, including any further data |
| yet perhaps the most important aspect is often | | | | investigation needed, and actions to progress your |
| overlooked: all your checks and analysis will be of | | | | list of issues. |
| little value if not actually put to some use. | | | | - Schedule to review progress and refresh the |
| While profiling your data, you should draw up an | | | | issue list. |
| action plan to take your data analysis further and | | | | Implicit in this plan is that concept that the data |
| to generate a real return on your data profiling | | | | profiling should not be considered a one-off task. |
| exercise. | | | | You will want to reuse elements of your original |
| The action plan should cover the following areas: | | | | analysis to keep your issues list up to date, and |
| - Reporting of results to interested parties. | | | | to ensure that new issues are not being |
| - Documentation of the source of the data, | | | | introduced. And of course there may be areas |
| checks undertaken, and any assumptions. | | | | which you want to investigate further. |
| - Any verified properties of the data. | | | | All too often, your data profiling exercise is |
| - Provide a prioritised list of issues, or at least a | | | | completed and everyone moves on to the next |
| list of issues and a mechanism to prioritise these | | | | stage of the project. Instead you should consider |
| later. | | | | a regular refresh of at least the key components |
| - Flag up any discoveries which directly impact the | | | | of your data profile, and perhaps take it forward |
| data project. Or which need further investigation. | | | | in to a more comprehensive data quality strategy. |
| - Flag up any issues which directly affect the | | | | |