Final Project Requirements

DATA 301

Introduction

This section should:

  • introduce the broader data context (social, cultural, historical, institutional)
  • introduce the data being analyzed (material context)
  • introduce the overarching research questions

Methods

This section should discuss the methods you used to process the data and the models you chose. Specifically, you should discuss:

  • How you cleaned / modified the original data and why.
  • What features you considered for your analysis and why.
  • What model(s) you chose for your analysis and why.

Visualizations

In this section you should present two (not three, not four) visualizations. The first visualization should address your primary research question and the second visualization should address your secondary research question (yes, only one!).

Additional criteria:

  • At least one plot must include 3+ variables.
  • Both plots must use non-default colors.
  • Both plots should follow the visualization “best practice” outlined in the course slides.

Model

In this section you need to fit a model(s) to the data to address your primary and secondary research questions. In this section you should explicitly state the modeling choices you made and the results of the model(s) you fit.

The following are minimum criteria for the model(s) you fit to your data:

  • Include a model covered after the midterm exam (KNN, logistic regression, linear regression, decision tree, k-means).
  • Use cross-validation to estimate the testing error of the model
  • Choose a “meaningful” model metric to summarize the fit of a model

Results / Implications

In this section you should connect the findings from your “Model” section to the primary and secondary research questions you posed at the beginning of the project.

Ethics & Limitations

This section should discuss the ethics of the model(s) that were used in your analysis and how it(they) could detrimentally impact people. This section should also outline the limitations with the model(s) that were fit. Based on how the data were collected, what population can your model be inferred onto?