6  Assignment 6: ANOVA and Payment Patterns in NYC Camera Violations

Statistical Significance vs Practical Significance

7 Overview

You are working as a data scientist hired by a law firm that specializes in fighting parking and camera tickets.

The firm wants to understand whether certain groups tend to pay more than others. Previously, we explored patterns by day of week and time of day. Now, you will conduct a structured ANOVA investigation using NYC camera violation data.

Your job is to determine whether payment differences across groups are:

  • Statistically significant
  • Practically meaningful
  • Strategically useful for marketing

Be sure to include a hyperlink somewhere in your introduction to the dataset via the
NYC Open Data Portal.

8 Learning Objectives

By completing this assignment, you will be able to:

  • Load public NYC data using an R package
  • Clean and prepare real-world categorical and numeric variables
  • Engineer time-based features (day of week, time of day)
  • Conduct multiple one-way ANOVA tests
  • Interpret F-values, p-values, and PRE
  • Distinguish between statistical and practical significance
  • Communicate inferential findings clearly in a reproducible report

9 Textbook Connection

This assignment builds directly from Chapter 5: Comparing Multiple Groups in Reproducible Research Using R.

Students are encouraged to review the chapter before beginning this assignment, as it provides the conceptual foundation and reproducible workflow demonstrated here.

10 Submission Instructions

Submit two files:

  1. Your .qmd file
  2. Your knitted .html file

Your document must:

  • Knit without errors
  • Contain clearly labeled sections
  • Include visualizations, descriptive statistics, and ANOVA results
  • Include written interpretation after each major section
  • Use at least one inline R expression

11 Assignment Tasks

You must complete the following sections in order.

11.1 1. Data Ingest

Load the NYC camera violations dataset using the nycOpenData package.

Your section must:

  • Clearly state what the dataset represents
  • Report how many rows are in the dataset (use an inline R expression)
  • Confirm that the dataset loaded successfully

Do not hard-code numbers in your text.

11.2 2. Data Cleaning and Feature Engineering

You must perform the following cleaning steps. You are responsible for determining how to implement them.

11.2.1 2A. Numeric Conversion

Ensure all payment-related variables (such as payment_amount and fine_amount) are stored as numeric.

Include a short checkpoint confirming the structure is correct.

11.2.2 2B. Date Cleaning

The issue_date column may contain inconsistent formats.

You must:

  • Retain only properly formatted dates
  • Convert issue_date into a Date object
  • Create a new variable called day_of_week

Explain briefly why cleaning dates is necessary before modeling.

11.2.3 2C. Time Cleaning

Using the violation_time column:

  • Convert it into a format that can be interpreted numerically
  • Create a new categorical variable called time_of_day

Your categories must include:

  • Morning
  • Afternoon
  • Night

Include a checkpoint showing how many observations fall into each category.

Explain why feature engineering improves interpretability.

11.3 3. Analysis 1: Day of Week and Fine Amount

You will now investigate whether average fine_amount differs by day_of_week.

Your section must include:

  • A grouped descriptive summary (means displayed in descending order)
  • A one-way ANOVA
  • A supernova() table
  • A post-hoc test (Tukey)

Then write a short paragraph answering:

  • Is the effect statistically significant?
  • How much variance is explained (PRE)?
  • Is the effect practically meaningful?
  • Why might statistical significance be misleading here?

11.4 4. Analysis 2: Time of Day and Fine Amount

Repeat the same structure, but now examine fine_amount by time_of_day.

Include:

  • Visualization (boxplot required)
  • Descriptive statistics
  • ANOVA
  • PRE interpretation

In your paragraph, explain:

  • Whether differences are meaningful
  • Why fine amounts may show significance even if policy does not change

11.5 5. Analysis 3: Violation Type and Fine Amount

Examine fine_amount by violation type.

Your section must include:

  • Descriptive statistics
  • ANOVA
  • PRE interpretation

In your paragraph, explain:

  • Why this model is statistically strong
  • Why it may not be conceptually interesting
  • What this teaches us about modeling fixed policy variables

11.6 6. Analysis 4: Violation Type and Payment Amount

Now shift to the more behaviorally meaningful variable: payment_amount.

Examine payment_amount by violation type.

Your section must include:

  • Visualization (boxplot required)
  • Descriptive statistics
  • ANOVA
  • PRE interpretation

In your paragraph, explain:

  • Why this analysis is more practically meaningful than fine_amount
  • What the size of PRE suggests
  • How a law firm might use this insight strategically

11.7 Final Reflection

In 5–7 sentences, answer:

  1. What is the difference between statistical significance and practical significance?
  2. Which of the four analyses above do you think is most useful for real-world decision-making?
  3. If you were advising the law firm, what is one concrete recommendation you would make based on your findings?

12 Reproducibility Practice

This assignment focuses on reproducible statistical modeling.

Your document must:

  • Use clearly labeled code chunks
  • Avoid hard-coding values in your narrative
  • Use inline R expressions where appropriate
  • Include at least one data-checking “checkpoint” in each major section
  • Render cleanly to HTML without manual intervention

The goal is that another analyst could rerun your entire workflow and verify every statistical conclusion.

13 Publishing Instructions

  1. Render your .qmd file to HTML.
  2. Confirm that warnings and messages do not appear in the final document.
  3. Submit both your .qmd file and your published .html report.