Study Notes: Data Analytics

Henley Business School, University of Reading
Course Objective: Learn to manage, visualize, analyze data sets, and infer meaningful conclusions.

Essential Processes for Data Analysts
- Descriptive Analytics: Analyzing past data.
- Predictive Analytics: Predicting future events using historical data.
- Prescriptive Analytics: Using data to alter outcomes.
Scope: The module covers foundational aspects of all three analytics processes.

Data Analytics: A field that bridges statistics and software development to address data-related issues.
- Key Steps:
  - Collect data
  - Organize and store it
  - Analyze it
  - Draw conclusions from findings
- Outcome: Produces data-driven recommendations.
Business Data Analytics: Focuses on addressing business-related questions utilizing data analytics principles.

Reasons for Importance: The importance of data analytics spans various industries and operations; specific reasons will be expanded in detailed notes.

The module does not cover every facet of data analytics, as the field evolves continually.
- Learning Outcome: Students should gain the competence and confidence to analyze data and perform independent research on methodologies not included in the course.

Types of Assessment:
- Multiple Choice In-Class Quiz (Week 7) - 50% of grade
- Data Analytics Project (Hybrid Online Assessment/Test) - 50% of grade
Recommendations: Attend all quizzes, workshops, and seminars; seek help early if needed.

Focus: Assessments focus on seminar and workshop material and practical applications rather than lecture content.
- Importance: Maintaining pace with seminar and workshop activities is critical for success.

Required Tools: R-commander
Optional Tools: Excel
- Compatibility: Available on both Windows and Mac, accessible via apps.
- Installation support: provided by Teaching Assistants during workshops.

Descriptive Statistics: Overview and importance discussed.
Caution Against Generalizing:
- Descriptive statistics should not be used to generalize findings; limitations will be explained through practical examples.

Scenario Setup: Mr. Rogers tests two feeds on chicken groups to assess which leads to higher weight gains to evaluate profit potential.
- Variables:
  - Independent Variable: Type of feed (Feed A and Feed B)
  - Dependent Variable: Weight of chickens
- Formal Hypothesis:
  - Null Hypothesis (H0): No difference in mean weights between groups
  - Alternative Hypothesis (H1): Difference exists in mean weights.
Data Access: Available as Roger.csv on Blackboard, part of the ChickWeight dataset.

Findings: Average weights show Feed A (142.95g) appears superior to Feed B (135.27g); however, this is inconclusive.

Method: Mr. Rogers splits chickens fed on Feed A into two groups for further weight measurement.
- Results: Group A1 has a mean weight of 149.62g, while Group A2 shows 136g; the differences exemplify natural variation rather than feed effect.

Key Takeaway:
- Certainty of findings is unattainable due to chance factors; instead, a probability % can indicate confidence in experimental outcomes.
Statements about significance reflect confidence (e.g., 90%, 95%, etc.) that differences observed are not due to chance.

Utility of Descriptive Stats:
- Summarizing data
- Visualizing trends
- Inputs for inferential statistics and modeling processes.

Parameter (Population):
- Example: Average Weight of Cross: 125g
- Average Price: £1.25
Statistic (Sample):
- Example: Average Weight: 130.4g
- Average Price: £1.44
- Probability related to star (P(star)): figures exemplified.

Sampling types covered:
- Simple Random Sample
- Systematic Sample
- Stratified Sample
- Cluster Sample

Qualitative Data:
- Nominal: No order (e.g., eye color)
- Ordinal: With order (e.g., Likert scale satisfaction levels)
Quantitative Data:
- Discrete: Whole numbers (1,2,3…)
- Continuous:
  - Interval: Lacks absolute zero (e.g., temperature in Celsius)
  - Ratio: Has absolute zero representing absence (e.g., weight).

Mean: Average value calculated using all data points.
- Formulation: $\bar{x} = rac{ ext{Sum of all observations}}{n}$
Median: The middle value when data is sorted. Not affected by outliers.
Mode: Most frequent value; applicable in qualitative data analysis.

Variance: Indicates how far data points spread from the mean.
Standard Deviation (SD): Measure of the amount of variation in a set of values.
- Formulations:
  - Population Variance: $<br>u^2 = rac{1}{N} imes ext{Sum of } (x_i - \bar{x})^2$
  - Sample Variance: $s^2 = rac{1}{n-1} imes ext{Sum of } (x_i - \bar{x})^2$
Importance of SD: Vital for assessing distance from the mean and determining confidence levels pertaining to the mean.

Context of IQ Distribution:
- Average IQ set at 100; standard deviation of 15.
- Representation of IQ follow a normal distribution indicated as: $IQ hicksim N(100, 15)$ .

Analytics Types: Descriptive, Predictive, and Prescriptive
Business Applications:
- Utilizing data for customer behavior insights, classification, security, market analysis, forecasting, etc.
Descriptive Stats Value: Important for summarization, visualization, and foundational input into inferential statistics; caution in generalization urged.
Central Tendency Methods: Mean (outlier susceptible), Median (robust), Mode (qualitative relevant).
Broad Data Types: Continuous, interval, ordinal.
Dispersion Measures: Critical for understanding variance impact on mean confidence.