Raw data and classification

Data, as first collected, are raw (unclassified). Picture a shop's sale bin where everything has been dumped together, or a kabadiwallah's heap of old things: finding what you need takes ages. A long list of 100 students' marks in no order is the same. Raw data are large and hard to use; it is difficult to draw any conclusion from them.

Classification arranges data into groups or classes according to some criterion. Then each item can be located easily, groups can be compared, and the data can be analysed. A good classification is clear and suits its purpose.

Data can be classified in four ways:

• Chronological: by time. India's population in each census year, from 1951 to 2011, arranged year by year, forms a time series.
• Spatial: by place. The yield of wheat in different countries, or rainfall in different states.
• Qualitative: by an attribute that cannot be measured, such as gender, religion, literacy or caste. People are sorted by whether they have the attribute or not, and can then be divided further: first into male and female, then each group into married and unmarried.
• Quantitative: by a characteristic that can be measured, such as marks, income, height or weight, grouped into classes (0-10 marks, 10-20 marks, and so on).