3  Classification and Tabulation of Data

Raw data, as it is collected, is almost always a disorganized jumble of numbers and labels with no visible pattern. Before any meaningful analysis can begin, that raw data has to be arranged, classified into meaningful groups, and tabulated into a clean table. This chapter covers exactly that process: turning scattered observations into an organized, readable structure.

3.1 Raw Data and Arrays

  • Raw Data: Data exactly as it was collected, unorganized and in no particular order. Also called ungrouped data.
  • Array: Raw data rearranged in ascending or descending order of magnitude. An array is the simplest possible organization of numerical data, and it is usually the first step before any further classification.

Example

The marks scored by 10 students in a quiz, as recorded by the invigilator, are the raw data:

42, 35, 58, 60, 35, 47, 52, 60, 38, 55

Arranging these values in ascending order produces the array:

35, 35, 38, 42, 47, 52, 55, 58, 60, 60

Notice how the array immediately makes it easier to spot the lowest score, the highest score, and repeated values, none of which were obvious in the raw list.

3.2 Classification of Data

Classification is the process of arranging data into groups or classes based on shared characteristics, so that observations within a group are similar to each other and different from observations in other groups. Classification is typically done on one of four bases.

  • Geographical (Spatial) Classification: Grouped by location, such as sales figures by state or city.
  • Chronological (Temporal) Classification: Grouped by time period, such as revenue by month or year.
  • Qualitative Classification: Grouped by an attribute or quality that cannot be measured numerically, such as gender or religion.
  • Quantitative Classification: Grouped by a measurable characteristic, such as age, income, or marks, usually organized into class intervals.
Basis What It Groups By Example
Geographical Location Sales by region: North, South, East, West
Chronological Time Monthly revenue for Jan–Dec 2025
Qualitative Attribute Employees grouped by gender or department
Quantitative Measurable magnitude Students grouped by marks: 0–20, 20–40, and so on

Real datasets are rarely classified on just one basis. A sales dataset might be classified geographically by region, chronologically by quarter, and quantitatively by revenue slab, all at the same time, which is exactly what a pivot table or a group_by() operation in R and Python lets you do in practice.

3.3 Tabulation of Data

Tabulation is the systematic presentation of classified data in rows and columns so that it can be read and compared at a glance. A well-constructed statistical table has several standard parts.

  • Table Number: A reference number for citing the table elsewhere in a report.
  • Title: A clear, concise statement of what the table shows.
  • Captions and Stubs: Column headings (captions) and row headings (stubs) that label what each part of the table represents.
  • Body: The actual numerical data, arranged in the cells formed by rows and columns.
  • Footnote and Source Note: Any clarifying notes about the data, and a citation of where the data came from.

Example

Table 1: Number of Students by Class and Gender, ABC School, 2025

Class Male Female Total
Class VIII 22 20 42
Class IX 25 23 48
Class X 20 24 44
Total 67 67 134

Source: School administration records, 2025.

A good statistical table follows a few simple rules: keep it as simple as the data allows, give it a clear and specific title, always mention the units of measurement, include row and column totals where they add value, and always cite the source. A table that needs a paragraph of explanation to be understood has usually failed at its job.

A common mistake is cramming too many variables into a single table, making it unreadable. If a table needs more than about five or six columns to make its point, it is usually better split into two focused tables or turned into a chart, which is exactly the territory the visualization chapters later in this book will cover.

3.4 From Arrays to Grouped Series

Once data has been classified and arranged, it can be organized into one of three kinds of statistical series, depending on how many distinct values are involved.

  • Individual Series: Each observation is listed separately, with no repetition or grouping, such as the array of 10 quiz marks above.
  • Discrete Series: Values are listed once each, alongside how many times each value occurs, appropriate when a variable takes only a limited number of distinct values, such as household size.
  • Continuous Series: Values are grouped into class intervals, or ranges, appropriate when a variable can take a very large number of possible values, such as income or marks out of 100.

Most real-world quantitative data, especially continuous data with many possible values, needs to be organized as a continuous series before it becomes readable. That grouping process, deciding how wide each class interval should be and how many classes to use, is the subject of the next chapter.


Summary

Concept Description
Raw Data and Arrays
Raw Data Data exactly as collected, unorganized and in no particular order
Array Raw data rearranged in ascending or descending order of magnitude
Bases of Classification
Classification Arranging data into groups so that similar observations sit together
Geographical Classification Classification of data by location, such as sales by region
Chronological Classification Classification of data by time period, such as revenue by month
Qualitative Classification Classification of data by a non-numeric attribute, such as gender
Quantitative Classification Classification of data by a measurable characteristic, often grouped into class intervals
Parts of a Table
Tabulation The systematic presentation of classified data in rows and columns
Table Number and Title The reference number and the clear, concise heading that identify a table
Captions and Stubs Column headings and row headings that label the parts of a table
Body of a Table The cells of a table that hold the actual numerical data
Footnote and Source Note Clarifying notes and the citation of where the tabulated data came from
Types of Series
Individual Series A series in which every observation is listed separately with no grouping
Discrete Series A series in which each distinct value is listed once alongside its count
Continuous Series A series in which values are grouped into class intervals rather than listed individually