data-anomaly-detector

What This Skill Does

The Data Anomaly Detector is a robust OpenClaw skill designed specifically for the construction industry to identify irregularities within complex project datasets. By leveraging statistical methods such as Interquartile Range (IQR) analysis alongside machine learning heuristics, this agent scans construction logs, financial spreadsheets, and schedule files to pinpoint inconsistencies. It flags potential issues ranging from fraudulent cost submissions and labor overcharges to critical schedule slippages and data entry errors that could otherwise escalate into project-wide delays. The skill standardizes anomaly detection across project lifecycles, ensuring that stakeholders receive actionable alerts with severity classifications and suggested resolutions.

Installation

To integrate this skill into your environment, run the following command via your terminal or OpenClaw interface:

clawhub install openclaw/skills/skills/datadrivenconstruction/data-anomaly-detector

Use Cases

• Cost Verification: Automatically audit procurement data against predefined industry benchmarks (e.g., concrete or steel pricing) to spot invoicing errors or budget overruns.
• Schedule Integrity: Analyze critical path method (CPM) schedules to detect illogical lags, unrealistic activity durations, or unexpected productivity drops among trade partners.
• Risk Mitigation: Identify recurring data anomalies that may signal failing infrastructure, underperforming subcontractors, or systemic project management flaws.
• Quality Assurance: Detect duplicate entries or impossible numeric values in resource management logs to maintain a clean source of truth for downstream reporting.

Example Prompts

1. "Scan the 'Q3_Project_Financials.csv' file and generate a report of all cost anomalies exceeding the standard threshold for steel and labor."
2. "Review the current master schedule for the high-rise project and highlight any activity durations that appear as statistical outliers or pattern breaks."
3. "Analyze the last month of equipment usage logs and provide a summary of anomalies with 'critical' or 'high' severity levels, including suggested corrective actions."

Tips & Limitations

• Predefined Thresholds: While the skill comes with built-in construction-specific thresholds for common materials, these should be calibrated to your specific regional market and project type for the highest accuracy.
• Data Quality: The effectiveness of the anomaly detection is directly tied to the consistency of your input formatting. Ensure column headers and date formats are clean before processing.
• Human Oversight: Always review the detected anomalies manually. The 'confidence' score provided by the agent is intended as a guide to prioritize review, not as an automated replacement for professional project controls oversight.