One of the most powerful methods of using ClustanGraphics is via an Excel spreadsheet. You can prepare and manage your data with Excel until it's in exactly the correct format, including using formulas to calculate your values, and then read it with ClustanGraphics in seconds.
Furthermore, when you have obtained your clustering results, simply copy and paste them back to your spreadsheet for further analysis - for example, in the spreadsheet below simply highlight column 7 and paste the cluster codes obtained from the cluster membership table that ClustanGraphics builds for all clustering procedures.
We need to say a few painful words here about speed. Several of ourdata mining users noticed that our Excel reader wasn't the fastest on the planet (apology below!). Well we've changed all that with release 5.04 (April 2001). For example, an Excel file of 40,000 cases and 5 variables that used to be excrutiating to read, now gets crunched in 3 seconds flat. Here's the test file, which fills a whopping 2.5MB of hard disk:
Reading this file into ClustanGraphics 5 now takes about 3 seconds on a Pentium III desktop PC.
What's more, you can optionally prepare your Excel files with a column of case labels, a row of variable names,missing values, mixed data types, and even a column of differential case weights.