General Social Survey (GSS) Dashboard

This repository contains data cleaning scripts, index construction, and visualization materials for my Tableau dashboard based on the General Social Survey (GSS).
It demonstrates skills in data preprocessing, index creation, reliability testing, and interactive visualization.

---

Getting Started

Prerequisites

Package	Version	Description
Python	3.10+	Primary language for preprocessing
pandas	≥1.5	Data manipulation and transformation
numpy	≥1.24	Numerical computing
pyreadstat	≥1.2	Read SAS-format GSS data files
pingouin	≥0.5	Reliability analysis (Cronbach’s α)
factor_analyzer	≥0.5	Exploratory factor analysis
scikit-learn	≥1.3	Standardization and scaling
matplotlib	≥3.7	Visual checks and diagnostics

Installation

# Clone this repository
git clone https://github.com/SammyShaw/GSS-Dashboard.git
cd GSS-Dashboard

# (Optional) create a virtual environment
python -m venv .venv
source .venv/bin/activate   # On Windows: .venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt


import pandas as pd
import pyreadstat
from sklearn.preprocessing import StandardScaler

# Read raw GSS SAS file (1972–2022 cumulative)
df, meta = pyreadstat.read_sas7bdat("gss7222_r3.sas7bdat")

# Select variables used in the dashboard
columns = [
    "YEAR", "AGE", "SEX", "EDUC", "DEGREE", "RACE", "REALINC", "PRESTG10",
    "HAPPY", "LIFE", "HAPMAR", "HAPCOHAB", "HEALTH", "RELITEN", "GOD", "BIBLE",
    "ATTEND", "PRAY", "TRUST", "HELPFUL", "FAIR", "HRS1", "SATJOB",
    "CONEDUC", "CONFED", "CONMEDIC", "CONARMY", "CONBUS", "CONCLERG",
    "CONFINAN", "CONJUDGE", "CONLABOR", "CONLEGIS", "CONPRESS", "CONSCI", "CONTV"
]
gss = df[columns].copy()

# Example: standardized "Happiness" index
happiness_vars = ["HAPPY", "LIFE", "HAPMAR", "HAPCOHAB"]
gss["happiness_index"] = gss[happiness_vars].mean(axis=1)
scaler = StandardScaler()
gss["happiness_index_z"] = scaler.fit_transform(gss[["happiness_index"]])

# Export a 10% sample (used in this repo)
sample = gss.sample(frac=0.10, random_state=42)
sample.to_csv("data/gss_z_sample.csv", index=False)

Summary & Context

This project demonstrates:

1. Visualization using Tableau
2. Preprocessing and analysis of a large public dataset

I chose this project to practice Tableau and to represent my background in the social sciences.
It also celebrates publicly available data and aims to make it more accessible.

The GSS, funded primarily by the National Science Foundation, has been collecting data on American political opinions and social behaviors since 1972.
While the analyses here show Americans’ declining trust in institutions alongside rising political polarization, the project also highlights the importance of maintaining federal support for longitudinal public data.

---

Variable Composition Table

Index / Measure	Type	Description	GSS Variables Included
Education	Single Measure	Years of school completed (standardized).	EDUC
Health	Single Measure	Self-reported health, 1 (poor) – 4 (excellent).	HEALTH
Happiness Index	Composite	Average of happiness, life excitement, and relationship happiness.	HAPPY, LIFE, HAPMAR, HAPCOHAB
Religiosity Index	Composite	Attendance, prayer frequency, importance of religion, and belief strength.	ATTEND, PRAY, RELITEN, GOD, BIBLE
Social Relationships Index	Composite	Time spent with family, friends, neighbors, and social outings.	SOCREL, SOCFREND, SOCOMMUN, SOCBAR
Social Attitudes Index	Composite	Interpersonal trust and perceptions of fairness/helpfulness.	TRUST, HELPFUL, FAIR
Work-Life Balance Index	Composite	Product of reversed work hours and job satisfaction for respondents working ≥10 hrs/week.	HRS1, SATJOB
Quality of Life Index	Composite	Average of Education, Health, Social Relationships, and Work-Life Balance indices.	Derived from other indices
Confidence in Institutions	Composite	Average confidence in 13 institutions (reversed coding).	CONEDUC, CONFED, CONMEDIC, CONARMY, CONBUS, CONCLERG, CONFINAN, CONJUDGE, CONLABOR, CONLEGIS, CONPRESS, CONSCI, CONTV
Confidence in Government	Composite	Confidence in federal government and congress.	CONFED, CONLEGIS
Confidence in Media	Composite	Confidence in television and the press.	CONPRESS, CONTV
Polarization	Aggregate	Standard deviation of selected opinion measures over time.	Derived (no raw columns)

---

Visualization

For the interactive version, visit my Tableau Public profile:*
50 Years of the General Social Survey *)

Key Skills Demonstrated

• Data cleaning and transformation (pandas, numpy)
• Reliability and factor analysis (pingouin, factor_analyzer)
• Index construction and standardization
• Data storytelling and Tableau dashboard design
• Large-scale data handling (70k+ rows, 6k+ columns)
• Documentation and portfolio presentation with Markdown and GitHub Pages

---

Reflections on Tableau Design

Tableau enables powerful data storytelling but requires tradeoffs between comprehensiveness, clarity, and interactivity.
While many dashboards focus on one storyline, this project intentionally embraces breadth — highlighting multiple social trends across 50 years of GSS data.

Key design choices:

• Centered trend lines to emphasize temporal change
• Color-coded index buttons with miniature sparklines
• Parameter-driven filters and calculated fields for modular control

Some technical limitations (e.g., pages vs filter interaction) prevented seamless toggling between “All Years” and yearly playback — a design constraint I plan to revisit in future versions.

---

Related Links

• Dashboard: 50 Years of the General Social Survey
• Portfolio: https://sammyshaw.github.io/
• Data Source: GSS NORC at the University of Chicago

---

Author: Samuel Shaw, PhD
Seattle, WA
LinkedIn