Quickstart.md

@@ -187,3 +187,24 @@ Level 4 is performed by the end-user based on the validation reports and execute
 
 ## Level 4 validation
 
+- Level 4 validation is done using the same validation reports obtained for levels 2 and 3, but using the regels tab
+
+<img src="uploads/6a291f196c728b9308e91c6138980fb8/Screenshot_2022-08-22_at_11.02.05.png" width="30%" />
+
+- This tab lists all validation rules whether they are violated or not.  The list is roughly sorted by importance
+
+<img src="(uploads/c4b72baa2076ad15cd91d4218f41cb6b/Screenshot_2022-08-22_at_11.03.24.png" width="60%" />
+
+- The first column shows the severity of violation of the rule and the second the rule itself.  Column 3 is discussed below.  Column 4 indicates how often a rule was violated and column 5 how large a percentage of the time the rule was satisfied (100% = good, 0% = bad).  The final column indicates which part of the model will have issues due to the violated model (ontology or indicator).
+
+- Violations with a low satisfaction percentage are likely problems with the data.
+
+- Column 3 indicates how often the validation rule was triggered.  This allows finding spuriously satisfied rules.  In this example, the first 8 rules are spuriously satisfied: no errors are found but that's not due to data adhering to the rule, it is because the dataset contained no data that matches the rule in the first place (a rule talking about time-limited contracts will never trigger on a dataset with only permanent contracts)
+
+<img src="(uploads/b9ee50dc225b1e820b3e329c8ec13d2d/Screenshot_2022-08-22_at_11.03.46.png" width="60%" />
+
+- Spuriously satisfied rules may or not be an issue with missing data: the data is not there, but depending on the organisation that may just be a reflection of reality
+
+- Rules at the very bottom with 100% satisfaction and a positive number of activations are likely no issue.
+
+# More about queries