Dit bericht verscheen eerder bij FOSSlife
One reason for this is that the right chart type to use depends heavily on both the actual nature of your data and on what, exactly, you want to see, learn, and tell others through that specific chart. It’s not uncommon to create multiple, wildly different charts from the very same set of numbers, each illustrating one distinct issue.
The other reason is that creating a chart can be harder than necessary if the columns and rows of data aren’t already all close to each other and in the right order when you first realize you may want to use them in a chart.
Never rush to the chart wizard when you need a chart. Instead, first draw a sketch of what your chart should look like on paper, trying different types until you are happy with the result. Then return to your spreadsheet. If data isn’t already grouped in the best way for drawing the chart you want, rearrange it accordingly.
You should, for example, recognize which data is actual raw numbers to plot and which data isn’t really numbers (as far as the chart is concerned, at least), but labels for data categories. One example of the last kind is a column that contains years of birth, which in the chart will be used only to classify and distinguish different age classes. If rearranging is not possible, however, don’t despair. I will show how to deal with such cases in a moment.
Correct grouping of data in the spreadsheet makes chart creation faster but may still tell little or nothing about which chart type should be chosen. The chart type to use depends on what kind of relationship there is between the data.
To understand this, consider a table with just two columns. If the relationship between the data in the two columns is biunivocal – that is (real mathematicians, forgive me for this approximation!), if each value in the first column corresponds to one and only one unique value in the second column (for example, the maximum temperature on each day of the current year) – then a line chart may very well be the best choice. If, instead, there is more than one value on the y-axis for any given x-axis value, a scatter graph would probably be much better.
This said, I strongly suggest only playing with line charts until you have become familiar with all the settings of the chart wizard. Afterwards, it will be much easier to master the other chart types and, above all, to make charts that communicate well, rather than merely trying to impress friends or colleagues.
The second and third panels of the wizard are where you configure the Data Range (Figure 5) and the Data Series (Figure 6), respectively, of your chart. The first term indicates the complete set of cells that contains all the data to plot or use as labels in the chart. A data series, instead, is one single subset of data inside that range, which generates one separate subplot in the chart.
Dit bericht verscheen eerder bij FOSSlife
