Pie chart or bar graph is usually used for categorical variables. Pie chart shows relative values, but it cannot be used, if the observation belongs to multiple categories. In Chart 1, it is presented education level of workers in one company. It is easily seen, that most of them have secondary level education. Only a few of them have no education after primary school. [br][br]

Bar graph enables cluster comparisons in a meaningful way. If there is a natural ordering between categories, it can be displayed on horizontal axis of the graph. In Chart 2, there is a clustered bar chart of education level by sex. It is easily seen, that males are higher educated than females. Their bars are getting higher and passing bars of females when education level is getting higher.

A discrete variable can either have  finite number of values  (1.1, 1.3, 1.5) or countably infinite number of single unequal values (like natural numbers). A value of continuous variable can be any real number (uncountably  infinite). [br][br]Frequency table describes the number of unequal values. For categorical and discrete numerical variables it means counting the number of observations for each unequal value. If a graph is used, it is usually a [i]histogram (histogrammi[/i]). Sometimes , there are used [i]percent frequencies [/i]and [i]relative frequencies[/i]. [br][br]Countably infinite variables and continuous values should be classified before counting frequencies. Usually, classifying loses specific information but describes general features more clearly. In Table 1 is presented heights of 25 school boy. You can change width of grouping with a slider and study, how it affects the histogram. [br][br]