Giuditta Parolini - The Hell, According to a Data Scientist | PyData Global 2023

Data scientists must recognize that data represents real people and human experiences, even when reduced to numbers - there are “tears behind the data”

Proper metadata and documentation are critical - datasets without sufficient metadata become effectively unusable

Standardization is essential when working with data:

• Use consistent country codes, currency codes, date/time formats
• Follow ISO standards
• Maintain proper CSV formatting
• Leave missing values empty rather than using placeholders

Machine readability should be a non-negotiable requirement when working with data

APIs should not be treated as magical solutions - they require proper expertise in web development and data engineering

AI should be approached critically:

• Avoid using AI indiscriminately without understanding the reasoning behind it
• Ensure AI is environmentally, economically and socially sustainable
• Recognize the fundamental differences between human intelligence and machine computation

Technical competence matters - people should not advocate for or implement solutions they don’t fully understand

Data work requires careful attention to context and human factors:

• Consider the real-world implications of the data
• Maintain interpersonal respect and understanding
• Remember that statistics represent human experiences

Data cleaning and standardization problems could often be avoided by following established standards and best practices

There’s a moral and intellectual imperative to handle data responsibly and ethically