Linking Frequency Tables with Exploratory Data Analysis (EDA)

Exploratory Data Analysis (EDA) is the first step in data analysis, aimed at understanding the structure, patterns, and anomalies in data before applying formal statistical models. A frequency table is one of the core tools of EDA.

How Frequency Tables Support EDA

1️⃣ Understanding Data Structure

EDA begins with knowing what the data looks like. Frequency tables summarise raw data into counts and classes, helping analysts quickly understand:

• Range

• Distribution

• Concentration of values

  
  

2️⃣ Detecting Patterns and Trends

EDA focuses on identifying patterns. Frequency tables reveal:

• Modal classes (where data are concentrated)

• Symmetry or skewness

• Seasonal or behavioural clustering (in time-series contexts)

  
  

3️⃣ Identifying Outliers and Anomalies

EDA aims to detect unusual observations early. Frequency tables help spot:

• Rare values

• Extreme observations

• Empty or sparse class intervals

  
  

4️⃣ Guiding Graphical Analysis

Frequency tables form the basis for EDA visual tools such as:

• Histograms

• Frequency polygons

• Ogives

These visuals provide deeper insights into distributional shape and spread.

5️⃣ Preparing for Statistical Modelling

EDA determines which statistical techniques are appropriate. Frequency tables help assess:

• Normality

• Need for data transformation

• Suitability of parametric vs non-parametric methods

  
  

6️⃣ Improving Data Quality

EDA emphasises data cleaning. Frequency tables help identify:

• Data entry errors

• Missing values

• Inconsistent measurements

  
  

Real-Life Example

Suppose, you collected the daily sales data (₹’000) of a Retail Electronics Store. The retail electronics store in Coimbatore recorded its daily sales (in ₹’000) for 45 working days. The following ungrouped data represent the daily sales figures.

Daily Sales (₹’000)

18.5 21.2         19.8         22.6         24.1         20.4         23.8         25.5         19.3         21.7

26.4         27.1         22.9         24.6         23.1         28.2         29.5         30.1         26.8         27.4

31.6          32.0       29.8         28.6         27.9        25.3         24.8         23.5         22.1         21.9

20.7         19.6         18.9         20.1          21.4 22.8         23.9         24.2         25.7         26.1

27.6         28.9         29.2         30.4         31.1

1. Find the minimum and maximum sales

2. Construct a frequency distribution

3. Draw a histogram

4. Calculate mean, median, and mode

5. Comment on the nature of the distribution

Step-by-Step Solution

Step 1: Find Minimum and Maximum Values

• Minimum sales = ₹18.5 thousand

• Maximum sales = ₹32.0 thousand

  
  

Step 2: Calculate Range

Range = 32.0 − 18.5 = 13.5

Step 3: Decide Number of Classes

Using Sturges’ Rule:

k = 1 + 3.322 log⁡10 (45)

≈1 + 3.322 (1.653)

≈6.5 ≈7 classes

Step 4: Compute Class Width

Class width = 13.57 ≈ 1.9

For convenience, take class width = 2

Step 5: Form Class Intervals with Class Boundaries

Class Interval (₹’000)

18 – 20

20 – 22

22 – 24

24 – 26

26 – 28

28 – 30

30 – 32

Step 6: Form Class Limits, if Required

Class Limits (₹’000)

18.1 – 20

20.1 – 22

22.1 – 24

24.1 – 26

26.1 – 28

28.1 – 30

30.1 – 32

Step 7: Tally and Frequency Distribution

Frequency Distribution of Daily Sales

Daily Sales (₹’000)	Frequency
18 – 20	5
20 – 22	8
22 – 24	8
24 – 26	7
26 – 28	7
28 – 30	6
30 – 32	4
Total	45

Step 8: Interpretation

• Most sales fall between ₹20,000 and ₹28,000. So, the store has stable mid-range daily sales. therefore, inventory and staffing can be planned around the ₹20–28k sales band

• A few days recorded very high sales above ₹30,000. The highest concentration is in ₹20–24 thousand. However, high sales days are occasional, likely due to promotions or festivals.

• The distribution shows mild positive skewness. This initial analysis provides a foundation for further EDA and decision-making.