Project

Spring 2018

Requirements

• You may use data from the internet, data that you’ve collected yourself, or data that you may have used for another class. However, this project is meant to be different than anything you’ve done before.

Skills

• RMarkdown
• Data cleaning and manipulation
• Data visualization
• Statistical analysis
• Interpretation of data visualizations and statisical analysis

Details

• 2 to 4 high quality plots. At least 1 of the plots needs to include at least 2 variables.
• A statistical analysis appropriate to answer your question of interest.
• You are encouraged to try new graphics or analyses that we didn’t cover in class!

Evaluation

• Overall presentation
• Visualization details
• Depth/appropriateness of visualization
• Statistical analysis details
• Depth/appropriateness of statistical analysis
• Accuracy of conclusions regarding visualization
• Accuracy of conclusions regarding statistical analysis

Report

1. Introduction (1 page).

• What is your question of interest and why is it important
• What other work has been done?
• What do your data look like? Where do they come from? How were they obtained? What are the variable types and units? What were the steps you took in cleaning the data?

1. Descriptions of visualizations (page length will vary). A summary of each graphic should contain

• the graphic itself
• one paragraph (3-5 sentences) summarizing the key takeaways of graphic and any interesting features / associations / relationships / etc. that the reader should understand
• one paragraph (3-5 sentences) describing the tools used to create the graphic (e.g., ggplot features, other plotting tools, etc.), the coding techniques necessary to manipulate the data in order to create the graphic (e.g., subsetting, etc.), and any graphical choices that were made (e.g., bandwidth / bin width choices, etc.)

1. Description of the statistical analysis (page length will vary).

• the analysis itself
• one paragraph (3-5 sentences) describing the type of analysis, tools used to conduct the analysis (e.g., lm, etc.), the coding techniques necessary to manipulate the data in order to conduct the analysis (e.g., subsetting, etc.), and any choices that were made regarding missing data and complex survey weights (if applicable)
• one paragraph (3-5 sentences) summarizing the key takeaways of the analysis and any interesting features / associations / relationships / etc. that the reader should understand

1. Conclusions (no more than 1 page). Practical findings and conclusions. Weaknesses in design. Other variables that you’d want to look at?

This is a formal report. Your writing should be professional and free of spelling/grammar errors. Your graphs should be high-quality, with titles, axis labels, legends, etc.

Notes

• You should show all code for your technical report.
• Please cite your sources, either using traditional format (Lopez, 2013) or links to URLs. See the Markdown cheatsheet for how to use URLs in Markdown.
• Consider the project requirements when thinking about what data you’d like to use. If you have a data set with 2 or 3 variables, it is likely insufficient for meeting the requirements above. However, you can certainly merge existing data sets together.
• You should submit a single .Rmd file and a single .html file containing the report.