Applications - Database Quality

The measurement of Database Quality is perhaps the simplest application of Daisy.

Before trying to analyse your database, whether you are using Daisy or not, always check it's quality!

Once the initial Daisy Chart has been drawn, then apply limits to the nodes. Normally, with Daisy you are looking for common records and links, but with Database Quality, it is often more useful to look for those nodes with only a few records.

Daisy has deliberately been designed with many different mapping formats. These enable, all possible checks to be applied to your database.

Make use of the following tips and tricks, when analysing Database Quality.

1. Make extensive use of the various 'Test' mapping formats.

   These convert the field into 'True' or 'False'.

2. Always put histograms on the nodes.

   This will give a quick indication of the relative occurrence of each particular node.

3. Map the first character of a field using the First Character mapping format.

   This will identify those that don't start with acceptable letters or numerics.

4. Map a numeric field, by Partition, Thousands, Millions or other mapping format, that will give about ten to twenty nodes.

   This should identify any stray numeric values.

5. Map a date field, by Day of the Week.

   This will give the spread of days over the week, which can be quickly checked.

6. Map a time field, by Hour.

   This will give the spread of time.

7. Aggregation in Daisy is also a powerful database checking tool.

   Use it to create totals, averages, maxima and minima for each node. These may well indicate errors in the database.

Daisy Analysis processes data for various clients.

One of the problems in one important application is that data is assembled from a variety of sources and must be checked and tested in as short a time as possible. A standard Daisy Chart is applied to the database to see that it follows the pattern of previous weeks.