- Home
- Skills
- Agents & Orchestration
- data insight agent
Works with the AI Agents you already use
data insight agent
Turn CSV/TSV data into statistical insights with NumPy-based cleaning, neural network modeling, and K-Means clustering.
$5
data insight agent
Example session with this skill installed
Analyze customer_data.csv to see what drives 'churn_status'. Use 5 clusters for segmentation.
- Read your context and instructions
- Compiled the data insight agent
- Generated the data export
I have completed the analysis of customer_data.csv using the target column churn_status. The process involved data cleaning, correlation mapping, training a neural network for feature importance, and K-Means++ clustering.
Data Analysis Report
| Metric | Result | Interpretation |
|---|---|---|
| Data Quality | 124 rows imputed, 12 outliers clipped | Minor quality issues in monthly_spend. |
| Model Performance | 0.24 R² (Moderate) | The features explain 24% of churn variance. |
| Top Driver 1 | tenure_months (Imp: 0.18) | Strongest predictor; shorter tenure increases churn risk. |
| Top Driver 2 | support_tickets (Imp: 0.12) | Higher ticket volume is associated with churn. |
| Correlation | -0.42 (Pearson) | tenure_months has a strong negative correlation with churn. |
| Segments | 5 Clusters detected | Cluster 2 (High-Spend) shows the lowest churn rate. |
Key Findings
Data Quality: 5% of the monthly_spend column contained missing values which were imputed. The script detected and clipped 12 outliers in the data_usage column to prevent model skew.
2.
Drivers: The model achieved a moderate predictive power (R² = 0.24). tenure_months is the primary driver; churned customers average 14.2 months vs 38.5 months for retained ones.
3.
Segments: Cluster 0 (n=450) is distinguished by high support_tickets (z = +2.1) and represents your highest risk group. Cluster 2 (n=890) is characterized by long tenure and high spend.
Next steps
- Review
data-insight-output/cleaned.csvfor the processed dataset. - Investigate the high ticket volume in Cluster 0 to address service friction.
- Run the analysis again with
--no-clipif you believe the spend outliers are legitimate high-value targets.
customer_data.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Raw CSV data often hides critical patterns behind noise, missing values, and outliers. This skill provides a statistically grounded analysis engine that cleans, profiles, and models your local datasets without sending data to external APIs.
What it does
- Data Quality Audit identifies and repairs types, handles missing values via imputation, and detects IQR outliers.
- Statistical Profiling computes Pearson correlations and identifies identifier columns to prevent data leakage.
- Predictive Modeling trains a NumPy-based neural network to determine feature importance and calculate held-out R² scores.
- Cohort Segmentation applies K-Means++ to group data points and identifies distinguishing features for each cluster.
- Automated Cleaning generates a
cleaned.csvfile alongside a structured JSON report for downstream tasks.
How it works
- Target Identification involves specifying the CSV/TSV path and the target column you want to predict or explain.
- Analysis Execution runs the
analyze.pyscript to perform cleaning, correlation, modeling, and clustering. - Report Retrieval extracts results from the generated
report.jsonandreport.mdfiles. - Statistical Interpretation provides a summary of data health, key drivers, and segment characteristics based on computed z-scores.
Frameworks & tools
This skill requires Python 3.9+ and relies exclusively on NumPy and Pandas. It operates entirely offline for data privacy.
Why this beats prompting it yourself
Generic LLM prompts often hallucinate correlations or fail to handle large datasets accurately. This skill uses deterministic NumPy implementations for all math, ensuring that every insight is backed by a verifiable computation rather than probabilistic guessing.
Use cases
- Churn Analysis to find which customer behaviors correlate most strongly with account cancellation.
- Lead Scoring to identify distinguishing features of high-value segments using K-Means++.
- Data Sanitization to automatically handle outliers and missing values before training external models.
- Feature Discovery to rank column importance using permutation importance on a held-out test set.
Known limitations
Large files exceeding 50,000 rows are automatically sampled for modeling and clustering. Excel files must be converted to CSV or TSV before processing.
Large files (>50k rows) are sampled. Excel files require manual conversion to CSV/TSV. Requires Python 3.9+ with numpy/pandas installed.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 10 days ago
Requires