Hi ,
I would like to make 2 points :
1. Before you develop a chart , first take a good look at the data , and see whether you can draw any conclusions ; this is after all the only reason we collect and analyse data viz. to see what conclusions we can draw , or what inferences we can make.
A chart is never merely a pictorial representation of data.
If this is accepted , then it follows that the whole purpose of a chart is to make the viewer accept your own conclusions / inferences , since the chart will be developed to project these conclusions / inferences. It is seldom that a chart viewer draws conclusions /inferences which were not intended by the chart developer ; if this is really so , then probably the chart is a badly developed one.
A first look at the data shows that there is not much variation , either within a series or between series.
In such cases , it always helps to draw an offset chart , which has the averages removed , and only the variations retained.
Attached is a sample. You can certainly improve on it.
It is immediately clear that there is a seasonal variation , which spikes around Christmas , New Year and probably the financial year end or something else in March. Whether this is really so or it is not , only you can say.
2. In the process of doing this , I saw that the cell formats are skewed so that negative values are not in thousands ; the custom cell format is :
_-* #,##0,_-;-* #,##0_-;_-* "-"??_-;_-@_-
If you see closely , the comma highlighted in red is missing in the negative part of the format string. I do not know whether this is intentional or not , but in any case , it is not correct.
Narayan
