Prerequisite(s): A numerical correlation analysis is defined and executed in the Profiling perspective of the studio.
To access a more detailed view of the analysis results of the procedure outlined in Creating a numerical correlation analysis, do the following:
Click the Analysis Results tab at the bottom of the analysis editor to open the corresponding view.
Click on Analysis Result to see more detail of the analysis results in the three different views: Graphics, Simple Statistics and Data.
Click Graphics, Simple Statistics or Data to show the generated graphic, the number of the analyzed records or the actual analyzed data respectively.
In the Graphics view, the data plotted in the bubble chart have different colors with the legend pointing out which color refers to which data.
The more the bubble is near the left axis the less confident we are in the average of the numeric column. For the selected bubble in the above example, the company name is missing and there are only two data records, hence the bubble is near the left axis. We cannot be confident about age average with only two records. When looking for data quality issues, these bubbles could indicate problematic values.
The bubbles near the top of the chart and those near the bottom of the chart may suggest data quality issues too, too big or too small age average in the above example.
From the generated graphic, you can:
clear the check box of the value(s) you want to hide in the bubble chart,
place the pointer on any of the bubbles to see the correlated data values at that position,
right-click any of the bubbles and select:
Show in full screen
open the generated graphic in a full screen
access a list of all analyzed rows in the selected column
The Simple Statistics view shows the number of the analyzed records falling in certain categories, including the number of rows, the number of distinct and unique values and the number of duplicates.
The Data view displays the actual analyzed data.
You can sort the data listed in the result table by simply clicking any column header in the table.