Marketing 338 - Exam 1

The Iceberg Principle

10% of the problem is seen and understood, 90% of the problem is not seen or understood

symptoms

visible or surface level issues

symptoms examples

company has declining sales

problem

underlying causes leading to symptoms

problem examples

poor marketing strategy, low quality products, ineffective customer service

primary data

data gathered by the researcher for a the specific purpose of addressing a research problem

primary data examples

survey, observation, experiment, interviews

secondary data

data previously collected from other sources that are already published

secondary data examples

census data, internet information, library database, books, journal articles

internal secondary database

data from within an organization

internal secondary database examples

Sales records, accounting information, customer database

external secondary database

Data from outside the organization

external secondary database examples

Industry reports, government publications, commercial databases, academic research papers

data preparation (3 steps)

integrate multiple sets of data

clean data

transform data

data integration

merges multiple data into one dataset for meaningful and valuable information

analyze data

Use appropriate analytical tools and techniques to derive insights from the data

report findings

Present the insights and recommendations in a clear and actionable manner.

relational database

A relational database is a collection of data organized into a set of tables. Each table contains data categories (variables) in columns and observations (cases) in rows

key variable

A variable common between two datasets, which can be used to link records between them.

vertical data integration

Combining datasets where each dataset contains different observations but shares the same structure (e.g., merging data from different time periods).

vertical integration excel command

append

horizontal data integration

Combining datasets where each dataset contains different variables but shares the same observations (e.g., merging data from different sources for the same time period)

horizontal integration excel command

Merge

=UPPER

converts text to uppercase

=PROPER

converts text to proper case (first letter of each word capitalized)

=TRIM

removes extra spaces from text

=Remove duplicates

deletes duplicate rows

text to colums

splits text into seperate colums based on delimiters

select right delimiters

Choose the correct delimiters (e.g., comma, semicolon) to separate text into columns.

incomplete data

Data that lacks attribute values or has missing attributes of interest, also known as Missing Data

examples of missing data

• Missing data as part of the data collection process.

• Non-responses on survey questions.

noisy data

data that contains errors or outliers

noisy data examples

• Salary = "-100" (error)

• Age = "222" (outlier)

• Errors due to human or computer mistakes during data entry

inconsistent data

data that contains discrepancies in codes or names

inconsistent data examples

• Date represented as "2/11/2019" in one row and "2019 Feb 11" in another.

• Different data sources leading to inconsistencies.

listwise deletion

Excludes the entire case (record) with missing data.

disadvantages (listwise deletion)

• Can be problematic if the missing items are not at random (NMAR).

• Reduces the power of the model due to a smaller sample size

advantages (listwise deletion)

• Can compare analyses because the sample is the same for different variables

pairwise deletion

Only analyzes cases with available data on each variable

advantages (pairwise deletion)

• Keeps as many cases as possible for each analysis.

• Uses all available information for each analysis.

disadvantages (pairwise deletion)

• Cannot compare analyses because the sample is different for different variables.

• Can lead to different sample sizes across analyses.

imputation

The process of replacing missing data with substituted values.

mean subsititution

Filling missing values with the mean of the observed values.

outliers

An observation that lies at an abnormal distance from other values in a random sample from a population.

wide format (unstacked)

Data presented with each different variable in a separate column.

long format (stacked)

Data presented with one column containing all the values and another column listing the context of the value

when to transform (long to wide)

If the datasets are merged without transformation, multiple identical rows for demographics are created.

when to transform (wide to long)

When summarizing wide-format datasets at a glance becomes difficult due to the number of variables.

transform long to wide (excel command)

pivot columns (convert rows to columns)

transform wide to long (excel command)

unpivot columns (convert columns to rows)

data visualization

the graphical representation of information and data. By using visual elements like charts, graphs, and maps, data visualization tools provide an accessible way to see and understand trends, outliers, and patterns in data.

qualitative data

Data that describes qualities or characteristics. It is non-numerical and can be observed but not measured. (Colors, textures, smells, tastes, appearances, beauty)

quantitative data

Data that can be measured and expressed numerically. It can be counted or compared on a numeric scale (Height, weight, length, price, temperature)

when to use a word cloud

• Analyzing customer feedback and reviews to identify common themes or issues.

• Monitoring social media discussions to gauge public sentiment about a brand or product.

• Summarizing survey responses to highlight frequently mentioned topics

Four elements for telling a story with data

understand the audience, decide key takeaways, identify the data, good criteria for data visualization

understand the audience

Know who your audience is and what they care about. Tailor your data presentation to meet their needs and expectations.

Decide Key Takeaways (Identify the Task)

Determine the main messages or insights you want to convey with your data. Focus on the most important findings that will help your audience make informed decisions

Identify the Data Type

Recognize the type of data you have (qualitative or quantitative) to select the appropriate visualization methods and analyses.

Good Criteria for Good Data Visualization

• Title and Labels: Ensure visuals have clear titles, labels, and axes information.

• Accurate Representation: Avoid misleading elements like non-zero baselines in bar graphs or incorrect proportions in pie charts.

• Simplicity: Remove unnecessary information or clutter to make the visuals easy to understand.

scales of measurement

The process of assigning numbers or labels to certain characteristics of objects according to pre-specified rules.

scale

A classification that describes the nature of information and the process of creating a continuum on which objects are located according to the amount of the measured characteristics they possess.

types of scales

nominal, ordinal, interval, ratio

nominal scale

Uses labels without any quantitative value. There is no ordering (Gender (Male, Female), Types of fruit (Apple, Banana, Cherry))

ordinal

• Ranks objects or arranges them in order, but does not indicate the magnitude of difference between them. ( T-shirt sizes (Small, Medium, Large), Rank in a competition (1st, 2nd, 3rd)

interval

• Scales with equal intervals between values, but no true zero point.(Temperature scales (Celsius, Fahrenheit), IQ scores)

ratio

Scales with a true zero point, allowing for comparisons of absolute magnitude (Height, Weight, Income, Age)

Why We Should Learn Scales of Measures

Choosing the right scales is crucial as they influence our decisions about which statistical methods to use. Different scales have different properties and applications.

why are descriptive analyses important?

summarize data to provide meaningful insights. They help understand the central tendency, dispersion, and overall distribution of data

mean

the average of a dataset

mode

the most frequently occurring value in a dataset

median

the middle value when the data is ordered

range

The difference between the highest and lowest values in a dataset

variance

The average squared deviation of each value from the mean.

standard deviation

The square root of the variance, indicating the spread of data around the mean.

Why Do We Care About Dispersion?

Understanding dispersion helps us assess the variability and consistency within a dataset. It complements measures of central tendency by providing a fuller picture of data distribution.

Relationship Between Measures of Central Tendency and Dispersion and Scales of Measures:

The choice of descriptive statistics depends on the scale of measurement. For instance, mean and standard deviation are appropriate for interval and ratio scales, while mode is suitable for nominal data.

Frequency

The number of times a particular value occurs in a dataset.

percentage frequency

The frequency expressed as a percentage of the total number of observations

bar chart

• Displays categorical data with rectangular bars representing different categories (Comparing discrete categories (e.g., favorite colors, types of pets))

histogram

Displays continuous data with adjacent bars representing frequency distribution (Showing the distribution of continuous data (e.g., age, income))

cross tabulation

A method to examine relationships between two categorical variables by creating a contingency table.

Side-by-Side Bar Chart

A bar chart that allows for the comparison of multiple categories side by side.

Segmented Bar Chart

A bar chart that displays stacked segments representing different sub-categories within each bar.

one variable

bar chart, histogram

two variables

cross tabulation, side by side bar chart, segmented bar chart

mode + scales of measure

nominal, ordinal, interval (if meaningful), ratio (if meaningful)

mean + scales of measure

interval, ratio

median + scales of measure

Ordinal (sometimes), interval, ratio

range + scales of measure

ordinal (sometimes), interval, ratio

variance + scales of measure

interval, ratio

standard deviation + scales of measure

interval, ratio

checklist for good data visualization

• titles, labels, x- and y- axis labels

• start with zero

• accurate proportions (pie chart)

• no unnecessary information or illustrations

when completing both vertical and horizontal integration, which do you complete first?

vertical (append) and then horizontal (merge)

inner join

left join (left outer join)

right join (right outer join)

Full Outer Join

Lefti Anti Join

Right Anti Join