Data Analyst

export default async function data_analyst(input) {
  console.log("🧠 Running skill: data-analyst");
  
  // TODO: implement actual logic for this skill
  return {
    message: "Skill 'data-analyst' executed successfully!",
    input
  };
}


{
  "name": "@ai-labs-claude-skills/data-analyst",
  "version": "1.0.0",
  "description": "Claude AI skill: data-analyst",
  "main": "index.js",
  "files": [
    "."
  ],
  "license": "MIT",
  "author": "AI Labs"
}

# Missing Value Imputation Methods Reference

This document provides detailed information about various imputation strategies and when to use them.

## Overview

Missing data is a common challenge in data analysis. The choice of imputation method significantly impacts analysis quality and should be based on:
- The type of data (numeric, categorical, temporal)
- The pattern of missingness (random, systematic)
- The percentage of missing values
- The relationship between variables

## Imputation Methods

### 1. Mean Imputation

**Description**: Replace missing values with the arithmetic mean of non-missing values.

**Best for**:
- Normally distributed numeric data
- Low to moderate missing rates (<20%)
- Variables without strong relationships to others

**Advantages**:
- Simple and fast
- Maintains sample size
- Preserves mean of the distribution

**Disadvantages**:
- Reduces variance
- Distorts correlations
- Not suitable for skewed distributions

**Example use case**: Imputing missing temperature readings, height measurements, or test scores.

---

### 2. Median Imputation

**Description**: Replace missing values with the median of non-missing values.

**Best for**:
- Skewed numeric distributions
- Data with outliers
- Ordinal data

**Advantages**:
- Robust to outliers
- Works well with skewed data
- Simple to implement

**Disadvantages**:
- Reduces variance
- May not preserve relationships between variables

**Example use case**: Imputing income data, house prices, or any right-skewed distribution.

---

### 3. Mode Imputation

**Description**: Replace missing values with the most frequent value (mode).

**Best for**:
- Categorical variables
- Binary variables
- Low-cardinality discrete variables

**Advantages**:
- Appropriate for categorical data
- Maintains most common pattern
- Simple interpretation

**Disadvantages**:
- May introduce bias if mode is not truly representative
- Reduces variability

**Example use case**: Imputing product categories, gender, yes/no responses.

---

### 4. Constant Value Imputation

**Description**: Replace missing values with a predefined constant (e.g., "Unknown", 0, -999).

**Best for**:
- High-cardinality categorical variables
- When missingness itself is informative
- Text fields

**Advantages**:
- Makes missingness explicit
- Useful for categorical analysis
- Simple to implement

**Disadvantages**:
- May create artificial category
- Not suitable for numeric analysis without transformation

**Example use case**: Imputing missing comments, optional survey fields, or product descriptions.

---

### 5. Forward Fill (LOCF - Last Observation Carried Forward)

**Description**: Replace missing values with the last observed value in sequence.

**Best for**:
- Time series data
- Sequential measurements
- Slowly changing variables

**Advantages**:
- Preserves temporal patterns
- Logical for continuous processes
- Maintains smooth transitions

**Disadvantages**:
- Assumes stability over time
- Can propagate measurement errors
- Not suitable for volatile data

**Example use case**: Imputing stock prices, sensor readings, or patient vital signs.

---

### 6. Backward Fill (NOCB - Next Observation Carried Backward)

**Description**: Replace missing values with the next observed value in sequence.

**Best for**:
- Time series with forward-looking data
- When future values are more relevant

**Advantages**:
- Useful for certain temporal patterns
- Complements forward fill

**Disadvantages**:
- Less intuitive than forward fill
- May not reflect actual pro