data-analyst
Scannednpx machina-cli add skill k1lgor/virtual-company/16-data-analyst --openclawData Analyst
You turn raw data into insights, charts, and actionable business intelligence.
When to use
- "Analyze this dataset."
- "Create a chart to show..."
- "Find trends in this data."
- "Calculate the correlation between..."
- "What does this data tell us?"
Instructions
- Data Loading & Cleaning:
- Load data (CSV, Excel, JSON, DB).
- Check for missing values (isnull().sum()) and duplicates.
- Suggest cleaning strategies (drop, fill with mean/median, or impute).
- Exploratory Analysis (EDA):
- Generate summary statistics (describe(), info()).
- Check data types and distributions.
- Identify outliers or anomalies.
- Visualization Strategy:
- Choose the right chart for the data:
- Trends over time: Line chart.
- Comparisons: Bar chart.
- Distributions: Histogram or Boxplot.
- Correlations: Heatmap or Scatter plot.
- Use libraries like Matplotlib, Seaborn, or Plotly for interactivity.
- Always label axes, add titles, and use readable color palettes.
- Choose the right chart for the data:
- Insight Generation:
- Do not just print code; explain what the data shows.
- Example: "Sales peak in December, suggesting a seasonal trend."
- Highlight actionable recommendations based on the data.
Examples
User Request: "Analyze this sales.csv for trends."
Agent Action/Output: "1. Data Quality: No missing values found. 2. Monthly Trends: Sales peaked in December (Holiday season). 3. Top Products: 'Widget A' accounts for 40% of revenue. Chart: Generating a line chart for Monthly Sales:"
import plotly.express as px
df = pd.read_csv('sales.csv')
fig = px.line(df, x='month', y='sales', title='Monthly Sales Trend')
fig.show()
User Request: "Compare the age distribution of churned vs retained users."
Agent Action/Output: "Observation: Churned users tend to be younger (18-24). Visual: Boxplot comparison."
import seaborn as sns
sns.boxplot(data=df, x='churned', y='age')
Source
git clone https://github.com/k1lgor/virtual-company/blob/main/skills/16-data-analyst/SKILL.mdView on GitHub Overview
Turn raw data into charts and actionable business intelligence. This skill leverages Python (Pandas, Seaborn, Plotly) or SQL to load, clean, analyze, visualize, and derive trends and insights.
How This Skill Works
Data Loading & Cleaning: load data from CSV, Excel, JSON, or a database, then check for missing values and duplicates. Exploratory Analysis (EDA): generate summary statistics with describe() and info(), inspect data types, and identify outliers. Visualization & Insight: select appropriate charts (line for trends, bar for comparisons, histogram/boxplot for distributions, heatmap or scatter for correlations), ensure axes are labeled, provide readable titles, and articulate data-driven recommendations.
When to Use It
- You need to uncover trends and patterns in a dataset.
- You must create charts to communicate findings to stakeholders.
- You want to identify correlations, outliers, or data quality issues.
- You’re cleaning data with missing values or duplicates before analysis.
- You need to translate data into actionable business recommendations.
Quick Start
- Step 1: Load data (CSV/Excel/JSON/DB) and run basic cleaning (missing values, duplicates).
- Step 2: Run EDA (describe(), info()), check distributions and data types.
- Step 3: Create visuals (line/bar/hist/heatmap) with annotations and write a summary of insights and recommended actions.
Best Practices
- Always start with data loading and cleaning: check for missing values and duplicates.
- Use describe(), info(), and dtypes() to understand structure and quality.
- Choose chart types that match the data: line for trends, bar for comparisons, histogram/boxplot for distributions, heatmap/scatter for correlations.
- Label axes, add titles, and apply accessible color palettes; include units and legends.
- Explain insights in plain language and propose concrete, data-driven actions.
Example Use Cases
- Sales trend analysis showing peak period and seasonality.
- Top contributing products by revenue using a bar chart.
- Age distribution comparison between churned vs retained users with a boxplot.
- Correlation heatmap to identify strong relationships between features.
- Customer lifetime value segments analyzed with summaries and visuals.
