The Quartz Guide to Bad data
FreeFreeFree tier
About The Quartz Guide to Bad data
The Quartz Guide to Bad Data is an exhaustive reference cataloging common problems encountered in real-world datasets, with practical suggestions for resolving them. Created by Quartz, the guide is organized by who is best equipped to solve each issue—the data source, the analyst, a domain expert, or a programmer. It covers topics such as missing values, duplicate rows, inconsistent date formats, data in PDFs, sampling bias, and p-hacking, among many others. The guide is freely available on GitHub under a Creative Commons license and welcomes community contributions via pull requests and translations.
Key Features
Exhaustive index of data quality issues with descriptions and solutions
Organized by problem-solving responsibility: source, analyst, expert, or programmer
Practical suggestions for when a problem cannot be fully resolved
Creative Commons licensed for non-commercial use
Community contributions via pull requests and translations into multiple languages
Pros & Cons
Pros
- Comprehensive coverage of a wide range of data problems
- Practical, actionable advice grounded in real-world reporting experience
- Freely available and open for community improvements and translations
- Well-organized structure makes it easy to find relevant issues
Cons
- Not a tool or software – provides guidance, not automation
- Primarily targeted at journalists and may not cover advanced data science techniques
- May not receive frequent updates (last activity not specified on the page)
Best For
Data journalism: identifying and documenting data issues before publicationData cleaning and quality assurance workflowsTeaching and learning about common data pitfallsReference for data analysts and scientists working with messy real-world data
FAQ
Is the guide free to use?
Yes, the guide is freely available under a Creative Commons Attribution-NonCommercial 4.0 International License.
Can I contribute to the guide?
Yes, pull requests are welcomed. There are also translations in Chinese, Japanese, Portuguese, Spanish, and partial Chinese; contributors can email Chris to have their translation added.