Experimenting With Sample Data

Graf-FX is a customised version of Microsoft Access that has been designed to help you analyse your databases in either a graffical fashion or using queries. The techniques that are employed are a particular database querying technique that is called data mining. Data mining is the use of a combination of cumulative queries and selective filtering to continually subset your data sets so that you can analyse trends in your data and look for data anomalies. For more advanced users of Access 95 and 97, this is similar to the use of filter by example except that this product allows you to graf the data while you filter it.

When Graf-fx starts up, you are presented with the main drilldown form and a introductory help form. A snapshot of that startup sequence is shown below.

Tip: To close the introductory form quickly, click anywhere on the form that shows the picture of the "cool graffer".

To initially experiment with graf-fx, try the sample data that comes with the database by clicking on the select pull down list.

Select the zSales_Demo table to use in this example of the products capabilities. Note that you can select from any table or query that Microsoft access has in this database including your own externally managed data. Once you select a table or query you will see that a whole lot of other buttons and fields then become visible. Graf-fx initially shows you only the options that are relevant to you current selections and reveals more controls as they become relevant.

Now you need to choose from the show as and group by selection boxes shown below.

Make two selections as follows.

Now you are ready to show grafs and queries of the Sum of All Sales categorised by the different Product Names.

Choose from the button showing the vertical graph bars to fire up the first consolidated graf or the yellow query button to fire up its associated query. The results of these are shown.

It is important to note that both the graf and the query can be reused in other Access applications. Both are opened as separate Access Windows and can be closed whenever you like by clicking on the cross in the top right corner of the window.

But what we are interested in is exploring the data, so to review why the Cola Sales are high, look at the second level drilldown area as shown below

Here we must select a subset filter from Cola, Lemonade or Orange and a Sub Group in the choice field. So we select Cola and choose a sub group of Region. This will show us all the Cola Sales by Region. Then by selecting the yellow question mark button on the second level, this produces the query results as shown. Selecting the histogram button will show the same results as a query.

From the query results we can see that the Sydney results are the highest, so we can choose to analyse the Sydney results by date. Try running the query or the graf.

But just showing Sum of Sales by date produces 20 different dates totals which is too many results for finding the time when high sales occur. So if you right click on the Sub Group By field, you will see a series of options that are shown above. Choose the FlashBuild option and select the show by calendar/quarter on the form that is shown.

This screen shot show the function that is returned by the flash builder. The resultant graf is shown below. Here the Sydney Coke sales have been grouped so that all sales are summed by a Quarter of a year.