Hack.Diversity Data Interview: Product Inclusion Analysis Challenge

Overview

This technical interview challenge will evaluate your skills in data cleaning, exploratory data analysis (EDA), feature engineering, and predictive modeling within a product inclusion context. Your task is to analyze user engagement with inclusive product features, focusing on accessibility trends and demographic impacts. The final submission should be in the form of a Jupyter Notebook or RMarkdown file, demonstrating your analytical skills and ability to extract insights.

Objectives

This challenge assesses the following skills:

• Data Cleaning: Address missing values and handle outliers in key columns.
• Exploratory Data Analysis: Analyze user engagement and feature adoption.
• Feature Engineering: Develop new features to enhance your analysis.
• Predictive Modeling: Build a basic model to understand user behavior.
• Version Control: Submit your work through GitHub, showcasing a clear commit history.

Preparation and Setup

1. Fork the Repository: Clone the provided repository. If no starter repo is provided, create a new one and commit your work frequently.
2. Make Meaningful Commits: Each major task (e.g., data cleaning, feature engineering, etc.) should have its own commit to track progress.

Instructions

1. Retrieve and Inspect the Dataset

• Dataset Access: You have a CSV file containing simulated user interaction data with inclusive product features.
• Columns: The dataset includes columns such as:
  • user_id: Unique identifier for each user.
  • feature_used: Type of feature accessed (e.g., "Colorblind Mode," "Multilingual," "Visual Aid").
  • session_duration: Length of the user’s session (in minutes).
  • engagement_score: Engagement level on a scale from 1 to 5.
  • accessibility_settings: Count of customizations enabled (0-5).
  • age_group: Categorizes users by age group.
  • gender: Gender (e.g., "Male," "Female," "Non-Binary").
  • additional_attributes: Additional user demographics (e.g., "LGBTQ+" or "Person of Color").

2. Data Cleaning & Preparation

• Missing Values: Identify any missing values and decide how to handle them, especially in session_duration and engagement_score.
• Outliers: Detect and manage outliers in session_duration and engagement_score.
• High Accessibility User Classification: Create a new column called high_accessibility_user, classifying users with three or more accessibility settings as a binary label (1/0 or True/False). Explain your choice and its implications for model interpretability.
• Intersectionality Analysis: Add a new column that combines age_group, gender, and additional_attributes to facilitate intersectional analysis. For example, label combinations as “Young Adult - Female - Person of Color.”

3. Exploratory Data Analysis (EDA)

• Engagement Score Analysis: Visualize engagement scores across features and demographics (e.g., age group, gender, additional attributes). Highlight which demographics show the highest engagement for each feature.
• Feature Usage by Demographic: Examine patterns in feature_used across demographic attributes. Use your new intersectional column to uncover trends that may not appear when demographics are analyzed individually.

4. Feature Engineering & Predictive Modeling

• New Features: Create features such as:
  • Average session_duration per feature_used.
  • Average engagement_score by accessibility_settings and demographics (e.g., age group, gender).
• Predictive Model: Build a simple predictive model (e.g., logistic regression or decision tree) to determine if a user is a high_accessibility_user. Use variables such as age_group, gender, additional_attributes, engagement_score, and feature_used.

5. Insights & Recommendations

• Key Findings Summary: Summarize insights for each analysis step. Identify which demographics show the most engagement with each accessibility feature.
• Actionable Recommendations: Provide data-driven recommendations on how the company can enhance the adoption of inclusive features, particularly among demographics with lower engagement levels.

Submission Requirements

• File Format: Submit a Jupyter Notebook or RMarkdown file containing code, visualizations, and a 3-4 sentence summary for each analysis section.
• GitHub Submission: Push your work to GitHub, making meaningful commits. Include comments for each major section.

Helpful Tips

• Exploratory Data Analysis: Use libraries like Matplotlib, Seaborn, or ggplot2 to clarify insights.
• Predictive Model: Keep it simple with logistic regression or a decision tree.
• Version Control: Use clear commit messages like “Completed data cleaning” or “Finished EDA.”

Evaluation Criteria

Your submission will be evaluated on the following:

• Data Cleaning & Preparation: Appropriateness and clarity of data handling.
• Insights: Meaningful analysis of trends in inclusive feature adoption.
• Modeling & Accuracy: Clarity in model choice and interpretation of results.
• Presentation: Clear explanations with supporting visualizations.

Good luck, and showcase your best analysis skills!