K-Graph 2

Reviewed by John S Davison


Issue 28

Jul/Aug 87

Next Article >>

<< Prev Article




Here's another of the K Series of application packages from Kuma, this one providing your ST with comprehensive business graphics and statistical analysis capability. Now in its Mk II version, it offers graphs and charts to suit most uses. It is GEM based and easy to use, so at last you have a simple way to turn all that boring numeric data into something a little more digestible.

Like other members of the K Series, K-GRAPH comes in the obligatory flimsy packaging, containing a single sided disk and slim, tutorial style instruction booklet. I found the booklet to be one of Kuma's better efforts, but for the price it could (should?) be a lot better.

K-GRAPH uses the standard Kuma installation procedure to produce a working copy from the supplied master disk. On a single drive system this requires about 30 disk swaps to complete – not as ridiculous as some of Kuma's other offerings, but irksome nevertheless.


The program produces graphs and charts from data provided either by existing data files from programs such as K-SPREAD or by keying it in through K-GRAPH's data editor. The usual data editing facilities are provided, so you can easily add, delete and change sets of data to be charted. Up to 20 data sets can be displayed on any one chart, and up to 80 across all chart types at any one time. A maximum of around 30000 data points can be held in memory on a 512K machine –so this is no toy program! In practice, though, charts get very messy with more than about six data sets displayed, so you'd probably never push the program to its theoretical capacity.

Data points can also be generated from a formula supplied by you, and K-GRAPH has a set of 18 built-in arithmetic and trigonometric functions for this purpose, together with conditional processing facilities. You just define the formula, provide the range and increment for X values, and K-GRAPH does the rest, at least, that's the theory – I had a lot of trouble getting it to work. I found the program's error messages unhelpful, and the instruction booklet inscrutable in this area. Kuma should have included a better explanation of the syntactical rules and practical examples to make this feature more easily usable.

A wide range of chart types can be produced from your data, such as line and scatter graphs, area charts, vertical and horizontal plain bar charts, stacked, overlaid, or three dimensional bar charts, and pie charts. You can flip at will between different chart types for a given set of data, this being as simple as clicking on the appropriate icon. Couldn't be easier.


Up to four windows can be open at any time, each displaying a different chart. Size and position of a chart within its window is easily changed under mouse control, and you can design your own graph line styles and fill patterns in terms of pixel layout and colour. Point markers may also be used on line graphs, there being a selection of shapes available for this purpose. They can be used without the connecting graph line, and in fact this is how you produce a scatter graph – it's really a line graph without the lines!

Scale points on the X-axis may be individually labelled, permitting the use of non-numeric items such as dates, month names and the like and there is also a general annotation feature which lets you place text anywhere on the graph in a variety of font styles and sizes. You can draw an arrowed line 7 connecting the text with any part of the chart to highlight something of special interest. The text block can be moved at any time, too, and if it has an arrow attached, the line I 'rubber-bands' so the head of the arrow remains pointing where it should as you reposition the text. Very neat.

A legend box showing which line or fill pattern represents which data is automatically generated for you. This is based on the names you give to the sets of data, and the lines/fill patterns you defined for use in plotting them. Again, this can be placed anywhere on the chart and can be repositioned at any time without damaging the underlying graphics. By clicking on one of the legends, you are cleverly taken straight into edit mode for the data it represents.

Further options allow you to overlay a grid on the two dimensional charts, to change X and Y origins for certain types of chart, and to choose whether pie charts display percentages or absolute data values against each slice. You can also choose to slide out any one slice to highlight it, although a bug in the review copy made this feature unusable.

If all this isn't enough, you can even save a K-GRAPH chart in DEGAS format and use that package to customise it even more. The possibilities seem endless!


When you've got the chart looking exactly as you want it, you can print it out upright or sideways, and enlarged and/or centralised in the X, Y or both directions. K-Graph displays a representation of the printed page, with a shaded area showing where your chart will print using the current settings. You can even slide the shaded area around using the mouse, for fine positioning.

It's designed to work with Epson printers. Unfortunately, there seemed to be a problem using it with my Star SG10 Epson compatible, as yet unsolved. I got round it by transferring the file to DEGAS and printed it perfectly from there.

One annoying feature I discovered with printing is that if you accidentally initiate it and you don't have a printer connected, the whole system locks up. The only way out is to reset and reload the program.

K-GRAPH can save data, labels, lines and patterns separately, allowing the flexibility of re-using existing patterns and labels in other charts. This can be done at any time – even when a chart is being displayed. Alternatively, you can choose to save data and labels together, or everything for the complete chart.

If needed, there are also a set of statistical functions, which work on a given set of data. These include mean, standard error, median, upper and lower quartiles, variance, standard deviation, sum of squares, skewness, and kurtosis. And how about the coefficient of variation and quartile coefficient of dispersion, for good measure? Then there are other functions, such as maximum, minimum, and count of data items, and slope and intercept of the line of least squares best fit based on the data points provided. The coefficient of correlation of data points to this line is also calculated.

Other functions require more than one set of data samples, and include paired and unpaired T-tests, F-test, Chi-Squared test, and regression analysis. I personally don't have much use for this sort of thing, but I'm sure there are people out there who use it all the time!


K-GRAPH works well, apart from the niggles mentioned earlier. If you need a general graphing, charting and statistical package for your ST, and can live with these problems, then go ahead and buy K-GRAPH. It's rather expensive purely for casual home use, but could be of great practical value for business, professional or educational applications. I enjoyed using it, and will continue to do so. What more can I say?