Project Assignment 3
Final RevealJS Presentation
Overview
For this final assignment, you will synthesize all your work into a polished RevealJS presentation that demonstrates your data analysis skills and communicates your findings effectively. Your presentation should tell a complete story from research question to policy implications, showcasing rigorous statistical analysis using your merged datasets.
This assignment builds on your research question and hypotheses (Assignment 1) and your cleaned, merged dataset (Assignment 2) to create a comprehensive analytical presentation.
Deliverable: A 7-10 minute RevealJS presentation that includes exploratory data analysis, statistical modeling, and policy recommendations.
Regression Modeling Requirements
Before creating your presentation, you must complete a full regression analysis using your cleaned dataset. This forms the analytical backbone of your presentation.
Model Selection and Specification
Choose your regression approach: - Linear regression (OLS): If your dependent variable is continuous - Logistic regression: If your dependent variable is binary/categorical - Other: Justify if a different approach is more appropriate
Write your regression equations using your actual variable names:
Main Model: \[ \text{[dependent\_variable]} = \beta_0 + \beta_1 \cdot \text{[independent\_var1]} + \beta_2 \cdot \text{[independent\_var2]} + \beta_3 \cdot \text{[control\_var]} + \epsilon \]
Alternative Models (you must fit at least 2-3 models): \[ \text{[Simpler or alternative specifications]} \]
Required Analysis Components
Your analysis must include all of the following elements, which will be presented in your slides:
- Exploratory Data Analysis (EDA)
- Distribution of key variables (histograms, boxplots)
- Correlation matrix or scatterplot matrix
- Identification of outliers or unusual patterns
- Data Transformations
- Document any variable transformations (log, standardization, etc.)
- Justify why transformations were necessary
- Show before/after distributions if relevant
- Model Fitting and Selection
- Fit multiple model specifications
- Compare models using appropriate criteria (R², AIC, etc.)
- Explain your final model selection process
- Regression Assumptions (if using OLS)
- Check linearity assumptions
- Examine residuals for heteroskedasticity
- Test for normality of residuals
- Address any assumption violations
- Results Visualization
- Create a professional regression table
- Generate coefficient plot with confidence intervals
- Include at least one substantive interpretation plot
- Predictions and Scenarios
- Use your model to make predictions for specific cases
- Show how changes in key variables affect outcomes
- Demonstrate practical application of your findings
Presentation Structure and Requirements
Your RevealJS presentation should include the following slides, integrating your statistical analysis throughout:
Minimum Required Slides:
Your presentation should include at least the following slides, but you may add additional slides as needed to tell your story effectively:
Slide 1: Title Slide - Project title, group member names, date
Slide 2: Research Question & Motivation - Clear statement of your research question - Why this question matters for policy/decision-making - Brief preview of your main finding
Slide 3: Data Overview - Description of your merged datasets - Key variables and their sources - Sample size and time coverage
Slides 4-5: Exploratory Data Analysis - Distribution of your dependent variable - Key relationships in the data (correlations, scatterplots) - Any interesting patterns or outliers discovered
Slide 6: Analytical Approach - Regression model type chosen (OLS/logistic) and why - Your main regression equation - Brief overview of model selection process
Slide 7: Model Selection & Assumptions - Comparison of alternative models - Key assumption checks (if OLS) - Justification for final model choice
Slide 8: Main Results - Professional regression table or coefficient plot - Highlight statistically significant findings - Effect sizes and practical significance
Slide 9: Model Application - Specific predictions or scenarios - “What if” analysis showing policy implications - Confidence intervals around predictions
Slide 10: Policy Implications & Conclusions - Direct answer to your research question - Actionable recommendations based on findings - Limitations and future directions
You may include additional slides if needed to adequately cover your analysis, but aim to keep your presentation focused and within the 7-10 minute time limit.
Presentation Guidelines
Technical Requirements: - Use RevealJS format (.qmd
file with format: revealjs
) - Include well-designed visualizations (ggplot2 recommended) - Ensure all plots are readable and professional - Include proper statistical notation where appropriate
Content Guidelines: - Focus on clear interpretation of statistical results - Emphasize practical significance, not just statistical significance - Connect findings back to your original hypotheses - Maintain focus on policy/decision-making applications
Presentation Tips: - Keep slide text minimal - let visuals tell the story - Practice timing to stay within 7-10 minutes - Prepare to explain your statistical choices if asked - Focus on insights, not just methods
Technical Checklist
Before submitting, ensure your analysis includes:
Submission
Please export your RevealJS presentation as a QMD file and submit it via Blackboard by the due date. Each group member should submit a copy so that we have something in Blackboard to grade. Your presentation should be self-contained and ready to present.
In addition to submitting your QMD file, you must also record your presentation using Zoom, Vimeo, or a similar screen recording tool. Your recording should be 7-10 minutes in length and should walk through your slides as if you were presenting to a live audience. Practice your presentation beforehand to ensure smooth delivery and proper timing.
Once you have recorded your presentation, upload it to a platform that allows sharing (such as Vimeo, YouTube, or Zoom cloud) and post the link to your recorded presentation in the Discord server. This will allow your classmates and instructor to view your presentations and provide feedback.
Please also be sure to export your presentation to Quarto pub and submit the link to the presentation in the Discord server.