Column Naming Conventions for ML/AI Projects¶

Overview¶

To maintain consistency and readability in our project, we use the "snake_case" naming convention for all column names in CSV, JSON, and other data files. This guide provides a set of rules and examples to help you name columns correctly, applicable to all types of data including structured, semi-structured, and unstructured data.

What is Snake_Case?¶

Snake_case is a naming convention where each word in a column name is lowercase and separated by underscores (_). This format is widely used in the Python ecosystem and ensures that column names are easily readable and consistent.

Rules for Snake_Case Column Naming¶

1. Use Lowercase Letters: All letters should be in lowercase.
2. Separate Words with Underscores: Use underscores (_) to separate words.
3. Be Descriptive: Column names should clearly indicate the contents or purpose of the column.
4. Avoid Special Characters and Spaces: Do not use spaces, hyphens, or any special characters other than underscores.

Examples¶

Structured Data (CSV File)¶

user_id, age, gender, purchase_amount, purchase_date
1, 34, "M", 45.67, "2023-05-12"
2, 29, "F", 78.90, "2023-05-13"

Semi-Structured Data (JSON File)¶

[
  {
    "user_id": 1,
    "age": 34,
    "gender": "M",
    "purchase_amount": 45.67,
    "purchase_date": "2023-05-12"
  },
  {
    "user_id": 2,
    "age": 29,
    "gender": "F",
    "purchase_amount": 78.90,
    "purchase_date": "2023-05-13"
  }
]

Unstructured Data (Text with Metadata)¶

{
  "documents": [
    {
      "doc_id": "12345",
      "title": "Analysis of Sales Data",
      "author": "John Doe",
      "content": "The sales data for Q1 shows a significant increase in revenue.",
      "timestamp": "2023-06-01T12:34:56Z"
    },
    {
      "doc_id": "12346",
      "title": "Customer Feedback Summary",
      "author": "Jane Smith",
      "content": "Customer feedback for Q1 has been overwhelmingly positive.",
      "timestamp": "2023-06-02T14:22:33Z"
    }
  ]
}

Common Column Names in ML/AI Projects¶

• user_id
• age
• gender
• purchase_amount
• purchase_date
• doc_id
• title
• author
• content
• timestamp

Tips¶

• Keep it Short and Simple: Aim for clarity but avoid overly long column names.
• Use Consistent Terminology: Use the same terms consistently across different columns (e.g., timestamp vs. date_time).
• Context-Specific Names: Tailor column names to the specific context and requirements of your project.

Common Mistakes to Avoid¶

• Uppercase Letters: UserID (Incorrect)
• Spaces: user id (Incorrect)
• Hyphens: user-id (Incorrect)
• Special Characters: user@id! (Incorrect)

By following these guidelines, we ensure that our column names are consistent, readable, and easy to work with in data processing and analysis tasks, regardless of the type of data.