The tidyviz Python Package

A Python package for tidying and visualizing survey data.

PyPI GitHub Docs

Overview

Survey researchers spend a disproportionate amount of time on repetitive data cleaning tasks—expanding multiple choice responses, validating answer ranges, flagging suspicious response patterns, and creating standardized visualizations. tidyviz streamlines this workflow by providing a unified toolkit for survey data preparation and visualization in Python.

I built this package to address common pain points I encountered working with survey data, where the same cleaning steps and visualization needs appear across nearly every project. Rather than writing custom scripts each time, tidyviz provides tested, reusable functions that handle these tasks consistently.

Key Features

The package is organized into two main modules:

Data Cleaning (tv.tidy)

• Expand and collapse multiple choice responses
• Validate response ranges and handle invalid entries
• Detect data quality issues (missing patterns, straight-lining, speeders)
• Check logical consistency with custom rules

Visualization (tv.viz)

• Single and multiple choice bar charts
• Survey-appropriate styling and color palettes
• Built-in percentage labels and sorting options
• Publication-ready output

Installation

pip install tidyviz

Example Usage

import pandas as pd
import tidyviz as tv

# Load survey data
df = pd.read_csv('survey.csv')

# Clean: Expand multiple choice responses
df_expanded = tv.tidy.expand_multiple_choice(df, 'favorite_colors')

# Validate: Check response ranges
df_clean, invalid = tv.tidy.check_response_range(
    df, 'satisfaction', min_val=1, max_val=5
)

# Visualize: Plot responses with custom styling
tv.viz.set_survey_style(palette='categorical')
tv.viz.plot_single_choice(
    df_clean, 'contact_method',
    title='Preferred Contact Method',
    show_percentages=True
)