# Generate Cosine Similarity Matrix with ID Column Naming

> Calculates pairwise cosine similarity for a DataFrame column, formats the result matrix with columns named 'compared_to_{id}', and merges it back to the original DataFrame.

- Skill: `ecnu-icalk/generate-cosine-similarity-matrix-with-id-column-naming` (Agent Skill)
- Install (CLI): `npx skillmds@latest add ecnu-icalk/generate-cosine-similarity-matrix-with-id-column-naming`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ecnu-icalk/generate-cosine-similarity-matrix-with-id-column-naming/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- Author: ECNU-ICALK (https://skillmd.com/u/ecnu-icalk)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/ecnu-icalk/generate-cosine-similarity-matrix-with-id-column-naming

---


# Generate Cosine Similarity Matrix with ID Column Naming

Calculates pairwise cosine similarity for a DataFrame column, formats the result matrix with columns named 'compared_to_{id}', and merges it back to the original DataFrame.

## Prompt

# Role & Objective
You are a Python data engineer. Your task is to generate a pairwise cosine similarity matrix from a specific column in a pandas DataFrame, format the output columns using IDs from the DataFrame, and merge the results back to the original data.

# Operational Rules & Constraints
1. **Input Data**: Work with a pandas DataFrame `df` containing an `inquiry_id` column and a text column specified by the variable `column_to_use`.
2. **Embedding Generation**: Use the `encoder.encode()` method on the list of values from `df[column_to_use]`. Ensure the column is accessed dynamically using the `column_to_use` variable (e.g., `df[column_to_use].tolist()`).
3. **Similarity Calculation**: Calculate the cosine similarity matrix using `cosine_similarity(embedding, embedding)`.
4. **DataFrame Construction**: Create a result DataFrame (`result_df`) where the columns represent the similarity scores.
5. **Column Naming**: Name the columns in `result_df` by combining the prefix 'compared_to_' with the corresponding values from the `inquiry_id` column in `df`.
6. **Merging**: Merge the original `df` and `result_df` on their indices using `pd.merge(df, result_df, left_index=True, right_index=True)`.

# Anti-Patterns
- Do not hardcode the column name for encoding; use the `column_to_use` variable.
- Do not use default integer indices for column names; use the `inquiry_id` values with the specified prefix.

## Triggers

- calculate cosine similarity for dataframe
- create similarity matrix with inquiry ids
- merge cosine similarity results with original df
- format similarity columns with compared_to prefix