COMM 1503 Types of Data

Data: Refers to the information collected for analysis.
- Population Data: Data that represents an entire population.
- Sample Data: Data drawn from a subset of the population.
Types of Data:
- Survey, Census, Parameter: concepts related to data gathering and analysis.
- Categorical Data: Data representing categories or distinct groups (e.g., types of drinks).
- Quantitative Data: Numerical data that can be measured (e.g., heights, ages).
Objective of the Class:
- Discuss the utility and summarization of data provided.
- Explore frequency distributions and their importance in analyzing data.

Definition of Frequency Distribution:
- A summary that shows all possible values in a dataset and their respective counts.
- Helps to identify how often each value appears.
Goals of a Frequency Distribution:
- Present every value that occurs in the dataset.
- Count how many times each value appears in the data.
Types of Frequency Distribution:
- For both numerical and categorical data, the frequency distribution is typically the first step in summarizing data.
Example of Frequency Distribution:
- If analyzing drinks sold, the distribution would show:
- What types of drinks were sold
- How many times each kind was sold.

Methods for Calculation:
- Calculations can be performed by hand or using Excel. Basic counting suffices for manual calculations.
Using Excel for Frequency Distribution:
- Function COUNTIF: Counts the number of cells that meet a specific criterion.
```
=COUNTIF(A2:B26, criteria)
```
- Using dollar signs ($) in formulas to fix certain values when dragging formulas down rows/columns to ensure consistent referencing in calculations.
Total Count and Frequency Analysis:
- The total frequency count should match the number of observations in the dataset, e.g., total drinks sold = 50.
Importance of Accuracy:
- Verify that the number of counts aligns with the actual data collected to maintain data integrity.

Relative Frequency:
- Represents the proportion of a total that a particular value contributes.
- Calculated as:
  $\text{Relative Frequency} = \frac{\text{Frequency}}{\text{Total Frequency}}$
Percent Frequency:
- Similar to relative frequency but expressed as a percentage.
- Calculated as:
  $\text{Percent Frequency} = \text{Relative Frequency} \times 100$
Summation of Frequencies:
- Absolute frequency totals should equal the total observations, while relative frequency should always sum to 1.

Defining Ranges for Bins:
- Rather than focusing on exact quantities (e.g., exact heights), the objective is often to classify data into ranges (bins).
Choosing the Number of Bins:
- Decisions on bin count (typically between 5 to 20) are stylistic and depend on the analyst's judgment.
Calculating Bin Width:
- Formula:
  $\text{Bin Width} = \frac{\text{Largest Value} - \text{Smallest Value}}{\text{Number of Bins}}$
- Example: For binning heights with a range of values, if the largest height is 73 inches and smallest is 58 inches, the width for 5 bins would be:
  $\frac{73 - 58}{5} = 3$
Understanding Edge Cases in Binning:
- Decisions must be made consistently about including or excluding boundary values when establishing bins to ensure non-overlapping categories.
- Use square brackets [ ] for inclusivity and round brackets ( ) for exclusivity in bin limits.

Definition of Histogram:
- A visual representation of frequency distribution where horizontal (x-axis) represents various bins while the vertical (y-axis) represents frequency, relative frequency, or percent frequency.
- Unlike bar charts used for categorical data, histograms represent continuous data without spaces between bars, indicating data continuity.
Shape of Histograms:
- Analysis of shape is crucial; look for symmetry or skewness.
- Definitions:
- Symmetrical Distribution: Both sides of the histogram mirror each other.
- Skewed Distribution: One side is longer than the other:
  - Right-Skew: Most data is clustered on the left, tail to the right.
  - Left-Skew: Most data is clustered on the right, tail to the left.

Cumulative Distribution Explained:
- Represents the accumulation of frequencies, tallying totals up to specific bin limits.
- Cumulative frequency is calculated as:
- For each bin, add the frequency of that bin to the cumulative total of the previous bins.
Applications of Cumulative Frequency:
- Particularly useful for determining how many observations fall below a certain value without concern for the precise distribution beyond that point.
Limitations:
- Cumulative distributions are not applicable for categorical data, as they rely on the assumption of order and continuity, which categorical data lacks.

Recap of Frequency Distributions:
- Apply concepts of absolute frequency, relative frequency, and percent frequency to analyze data.
- Histograms visualize distribution effectively, revealing insights about shape and skewness.
- Understand the role of cumulative frequency in summing observations to establish overall trends in data analysis.