outlier-detection-handler

What This Skill Does

The outlier-detection-handler is a specialized OpenClaw skill designed to bring rigorous statistical discipline to your data analysis workflows. Unlike general-purpose analysis scripts, this tool enforces structured execution, requiring you to define explicit assumptions and clear output boundaries before computation begins. It serves as a robust mechanism to identify, filter, and handle statistical outliers within datasets, ensuring that your final analysis is not skewed by noise or anomalies. The skill is packaged with a primary executable, scripts/main.py, which is built to support reproducible results and consistent, reviewable output formats. By mandating a documented fallback path for missing inputs or execution errors, it helps users maintain clean, audit-ready data pipelines.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/aipoch-ai/outlier-detection-handler

Ensure that you have Python 3.10+ installed and that the required dependencies (numpy, scipy) are satisfied in your current virtual environment, as defined in the repository's requirements.txt file.

Use Cases

• Research Data Cleaning: Quickly isolate and flag anomalous readings in large experimental datasets without manually sifting through raw values.
• Quality Assurance Pipelines: Automate the detection of faulty inputs in recurring log analysis tasks, providing a documented fallback for invalid data points.
• Statistical Reporting: Generate highly reproducible reports where the exclusion criteria for outliers is explicitly documented and consistent across multiple runs.

Example Prompts

1. "Run the outlier detection handler on /data/monthly_sensor_logs.csv, using a Z-score threshold of 3.0 and outputting results to /reports/analysis_v1.json."
2. "Can you help me identify outliers in my sales performance data using the outlier-detection-handler? I need to exclude values that fall outside the 1.5 IQR range."
3. "Execute the outlier-detection-handler for my current dataset, but configure it to prioritize preserving missing data entries as null instead of dropping them."

Tips & Limitations

To get the best results, always define your threshold criteria and scope limitations clearly before running the tool. Because this skill enforces strict boundaries, it may error out if input files do not match the expected schema. If the output appears incomplete, review the documented assumptions generated in the metadata output. Remember to use the python -m py_compile scripts/main.py command for a quick syntax validation before processing large datasets to avoid runtime interruptions.