Introduction to Data Representation and Frequency

TL;DR

You'll learn how to organize raw information into meaningful groups using frequency tables. These tables show how often each value or group appears, making large datasets easier to understand. This foundation is crucial for further statistical analysis and visualizing data effectively.

1. The Mental Model

Imagine you have a huge pile of individual Lego bricks. Data representation is like sorting those bricks by color, size, or shape into labeled bins, so you can easily see how many of each type you have. Frequency is simply counting how many bricks are in each bin.

2. The Core Material

When you collect data, it often comes in a raw, unorganized form. This can be hard to make sense of, especially with many observations. Data representation is all about organizing this raw data into a more understandable format. Frequency helps us count how often specific values or categories appear in our dataset.

Raw Data vs. Organized Data

Photo by Lukas Blazek on Pexels

Let's say you asked 20 people their favorite color, and you got these answers:
Red, Blue, Green, Red, Yellow, Blue, Red, Green, Blue, Red, Yellow, Green, Red, Blue, Blue, Green, Red, Blue, Green, Yellow

This is raw data. It's just a list. To make it useful, we need to organize it.

Frequency Distribution

Photo by wal_ 172619 on Pexels

A frequency distribution is a table that shows the count (frequency) of each distinct value or category in a dataset.

Here's how you'd build one for the favorite colors:

1. List all unique categories: Red, Blue, Green, Yellow.
2. Count how many times each category appears:
   • Red: 7
   • Blue: 7
   • Green: 5
   • Yellow: 3
3. Organize into a table:

Let's re-do the count for the original 20 items:
Red: 7
Blue: 7
Green: 5
Yellow: 3
Total: 7+7+5+3 = 22. Wait, the raw data list only has 20 entries.
Red, Blue, Green, Red, Yellow, Blue, Red, Green, Blue, Red, Yellow, Green, Red, Blue, Blue, Green, Red, Blue, Green, Yellow

Let's do it carefully:
Red: (1), (4), (7), (10), (13), (17) -> 6
Blue: (2), (6), (9), (14), (15), (18) -> 6
Green: (3), (8), (12), (16), (19) -> 5
Yellow: (5), (11), (20) -> 3
Total: 6 + 6 + 5 + 3 = 20. Perfect, now it matches!

So the correct table is:

Relative Frequency

Photo by Oussama Bergaoui on Pexels

Sometimes, knowing the raw count isn't enough; you might want to know the proportion or percentage of each category. This is called relative frequency.

Relative Frequency = (Frequency of Category) / (Total Number of Observations)

For our favorite colors:
* Red: 6 / 20 = 0.30 (or 30%)
* Blue: 6 / 20 = 0.30 (or 30%)
* Green: 5 / 20 = 0.25 (or 25%)
* Yellow: 3 / 20 = 0.15 (or 15%)

Grouped Frequency Distribution (for numerical data)

Photo by RDNE Stock project on Pexels

When you have a wide range of numerical data, listing every single value would make the table too long. Instead, you group the data into class intervals or bins.

Imagine you collected the heights (in cm) of 30 students:
155, 160, 162, 165, 158, 170, 171, 159, 163, 166, 172, 157, 161, 164, 168, 173, 156, 160, 162, 165, 169, 174, 158, 161, 163, 167, 170, 175, 159, 166

Here's a process for creating a grouped frequency distribution:

1. Find the range: (Highest value - Lowest value)
   Range = 175 - 155 = 20 cm
2. Decide on the number of classes (bins): A common rule of thumb is between 5 and 10 classes. Let's aim for 5 classes.
3. Calculate class width: Range / Number of classes
   Class Width = 20 / 5 = 4 cm
   It's often good to round up to an easy number. Let's use a class width of 5 cm to make intervals cleaner.
4. Determine class limits: Start from the lowest value or slightly below it, adding the class width to get the upper limit. Make sure classes are exhaustive (cover all data) and mutually exclusive (no overlap).

Original 30 heights:
155, 160, 162, 165, 158, 170, 171, 159, 163, 166, 172, 157, 161, 164, 168, 173, 156, 160, 162, 165, 169, 174, 158, 161, 163, 167, 170, 175, 159, 166

Let's tally each individually:
155-159: 155, 158, 159, 157, 156, 158, 159 (7)
160-164: 160, 162, 163, 161, 164, 160, 162, 161, 163 (9)
165-169: 165, 166, 168, 165, 169, 167, 166 (7)
170-174: 170, 171, 172, 173, 174, 170 (6)
175-179: 175 (1)
Total: 7+9+7+6+1 = 30. Finally correct!

Data Representation Process

The overall flow of dealing with raw data looks like this:

graph TD
    A["Raw Data (unorganized)"] --> B{"Categorical or Numerical?"}
    B -- "Categorical (e.g., colors)" --> C["List unique categories"]
    C --> D["Count frequency for each category"]
    D --> E["Create Frequency Table"]
    B -- "Numerical (e.g., heights)" --> F{"Wide Range?"}
    F -- "Yes" --> G["Group into Class Intervals (bins)"]
    G --> H["Count frequency for each interval"]
    H --> E
    F -- "No (small range)" --> C

3. Worked Example

Let's track the number of books read by 25 students in a month:
3, 5, 2, 8, 4, 3, 5, 6, 2, 7, 4, 3, 5, 6, 8, 4, 2, 5, 7, 3, 6, 4, 5, 3, 2

1. Identify unique values: 2, 3, 4, 5, 6, 7, 8

2. Count frequency for each:

   • 2: 5 times
   • 3: 5 times
   • 4: 4 times
   • 5: 6 times
   • 6: 3 times
   • 7: 2 times
   • 8: 2 times
     Total = 5+5+4+6+3+2+2 = 27. (Oops, the prompt says 25 students. Let's recount from the list exactly.)

   Books read:
   3, 5, 2, 8, 4,
   3, 5, 6, 2, 7,
   4, 3, 5, 6, 8,
   4, 2, 5, 7, 3,
   6, 4, 5, 3, 2

   • 2: (3rd, 9th, 17th, 25th) -> 4 times
   • 3: (1st, 6th, 12th, 20th, 24th) -> 5 times
   • 4: (5th, 11th, 16th, 22th) -> 4 times
   • 5: (2nd, 7th, 13th, 18th, 23rd) -> 5 times
   • 6: (8th, 14th, 21st) -> 3 times
   • 7: (10th, 19th) -> 2 times
   • 8: (4th, 15th) -> 2 times
     Total: 4+5+4+5+3+2+2 = 25. Perfect!

3. Create Frequency and Relative Frequency table: