[DAGE Saves the Day]

Group: DAGE Course: PSYCH 755, Summer 2026

Communication apprehension: How accurately can semantic features from participants’ open-ended responses predict their dominant type of communication apprehension?
Transportation use: How accurately can semantic features from participants’ descriptions of their ideal travel predict whether their transportation use is public-transit dominant, rideshare dominant, or has no dominant mode?

Members

Name	GitHub username
Grant Mooslin	grantmooslin
Emily Huffaker	emilyhuffaker
Andrew Sanford	aksanford
Dasey Dang	daseydang

Project Structure + Contents

Path	Description
index.qmd	The primary manuscript. Start here.
contributions.md	Who owned what.
memos/	Individual research memos, one per member.
references.bib	Shared BibTeX file for the manuscript and memos.
data_org/	Houses original/raw data for the project
data_clean/	Houses cleaned or otherwise modified data for the project
scripts/	Analysis scripts and neural network models
results/	Model outputs and training histories

Environment Setup

Prerequisites

• Python 3.11 (required for TensorFlow compatibility)
• R (for data cleaning and EDA scripts)
• Quarto (for rendering the manuscript)

Installing Python Dependencies

# Install Python packages from requirements.txt
pip3 install -r requirements.txt

The requirements.txt includes:

• numpy, pandas (data processing)
• scikit-learn (machine learning utilities)
• tensorflow (neural networks)
• matplotlib, seaborn (visualization)
• jupyter, ipython, pyyaml (notebook support)

Installing Quarto

Visit quarto.org to install Quarto for your operating system.

Data Setup

Before running any analysis scripts, place the following raw data files in data_org/:

• PRCAQualtricsExport_FileC.csv
• PRCAProlificExport_FileA.csv
• PRCAProlificExport_FileB.csv

The expected location for each file is:

DAGE/
└── data_org/
    ├── PRCAQualtricsExport_FileC.csv
    ├── PRCAProlificExport_FileA.csv
    └── PRCAProlificExport_FileB.csv

Download or copy these files from the original data source (e.g., Qualtrics/Prolific exports or the project's shared drive).

Important: Do not commit data files to git. The data_org/ directory should contain only the original raw data files.

Reproducing this project

Step 1: Clean the data

# Run the data cleaning script to generate cleaned datasets
quarto render scripts/eda/clean_data.qmd

This creates cleaned data files in data_clean/:

• survey_clean.csv
• survey_transportation.csv
• survey_apprehension.csv

Step 2: Render the manuscript

# Set the Python environment for Quarto
export RETICULATE_PYTHON=/Library/Frameworks/Python.framework/Versions/3.11/bin/python3

# Render the main manuscript
quarto render index.qmd

This will:

• Run the neural network models
• Generate figures and tables
• Create index.html in the project root

Alternative: Use system Python default

If your system Python is at /Library/Frameworks/Python.framework/Versions/3.11/bin/python3, you can add the reticulate setup directly to index.qmd (already included) and simply run:

quarto render index.qmd

To preview the manuscript while editing, run:

quarto preview index.qmd

Notes

• The neural network models use TensorFlow/Keras and may take several minutes to train during rendering
• Model outputs are cached in results/ to speed up subsequent renders
• The freeze: auto setting in index.qmd caches code execution results when source code hasn't changed

DAGE is life
DAGE stands for Data Analysis is Great and Excellent
Or Dasey, Andrew, Grant, Emily.