Governata examines column content across your sources and surfaces missing and duplicate values, non conforming formats, and distribution patterns, so quality rules are built on reality rather than assumption.
Content scanning
Missing and duplicate values
Non conforming formats
Distribution patterns
Every rule rests on an assumption about the shape of the data, and where that assumption is wrong the rule either fails or waves the error through.
Profiling comes before the rule, because a rule built on a wrong assumption produces wrong results with confidence.
Analyzing the actual values in each column to establish what it holds rather than what it is assumed to hold.
Showing the share of empty or absent values in each column, the most common quality problem of all.
Surfacing repeated values within a single column, particularly in columns expected to be unique.
Identifying values that deviate from the prevailing pattern, such as a different date or number format.
Showing how values are distributed and where they fall, so outliers appear before they affect any analysis.
Turning what profiling reveals into quality rules applied to the same source.
Choosing the data source or table whose content is to be examined.
Analyzing column content and extracting the characteristics of the values within it.
Reviewing the missing values, duplicates, deviating formats, and distribution patterns surfaced.
Turning the findings into a quality rule applied to the source, with its result measured.
Knowing the real state of the data before writing any rule, so rules rest on reality.
Understanding the shape and range of the data before analysis begins, rather than finding anomalies in the result.
Knowing what needs remediation before data is moved or merged with other sources.
A picture of source condition that supports prioritization in quality improvement programs.
The type the values genuinely hold, which may differ from the type declared in the table.
How many values are present and how many are empty, the first thing that determines whether a column is usable.
How many different values a column holds, an indicator of its nature and uniqueness.
The lowest and highest values, revealing anything outside a sensible range.
It is the examination of actual data content to extract its characteristics, such as completeness, uniqueness, and value patterns, before it is used or quality rules are written for it.
Profiling reveals what the data actually is. A rule defines what it should be. The first logically precedes the second.
Before building quality rules, when a new data source is added, and before any significant merge or migration.
Because an outlier may be an entry error or a genuine rare case. Profiling surfaces it, and review is what determines which it is.
No. Data changes continuously, and periodic profiling is what reveals deteriorating characteristics before their effect shows.
Profiling is the discovery stage. Measurement and remediation follow it, but both begin from knowing what the data actually holds.
Missing values, duplicates, deviating formats, and distribution patterns, surfaced per column before any quality rule is built.