Analyze Data and Generate Actionable Reports

Analytics Reporter is an autonomous agent that analyzes data files, extracts metrics, identifies trends and anomalies, and produces a full report with actionable recommendations. It works through six stages: data discovery (using Glob to find CSV, JSON, and log files, then Read to identify structure and catalog available metrics and dimensions), data validation (checking for missing values, duplicates, and outliers, verifying freshness and completeness, and documenting limitations), metric calculation (extracting KPIs, growth and conversion rates, statistical summaries such as mean, median, and percentiles, and correlations between metrics), trend analysis (comparing the current period against previous periods, identifying seasonal and cyclical patterns, detecting anomalies, and segmenting by relevant dimensions), insight generation (surfacing the top 3-5 findings, determining root causes, benchmarking against industry standards where possible, and prioritizing by business impact), and recommendation development (specific, actionable recommendations with implementation difficulty, expected impact, and monitoring strategies).

Use it when you have raw data files and need a structured report with statistically grounded insights rather than a raw dump of numbers. It is not a substitute for domain expertise on ambiguous findings: the agent's own guidelines call for validating assumptions and cross-referencing multiple data sources before trusting an anomaly, and confidence levels should be stated rather than implied.

Input is one or more data files (CSV, JSON, log files) discoverable via Glob. Output is a structured report: an executive summary (a 2-3 sentence overview, critical metrics at a glance, and a primary recommendation), a key metrics dashboard table (metric, current value, previous value, percent change, and status), detailed analysis (month-over-month and year-over-year trends, segment performance, anomaly detection, and correlation insights), recommendations split into high, medium, and low priority, and a data appendix documenting sources, timeframes, methodology, and known limitations.

### Growth rate calculation
awk '{if(NR==1) prev=$2; else curr=$2} END {print (curr-prev)/prev*100}' metrics.csv

### Moving average for trend smoothing
awk '{sum+=$1; count++; if(count>7){sum-=prev[count%7]}; prev[count%7]=$1; if(count>=7) print sum/7}' data.csv

Relies on file-discovery and file-reading tools (Glob, Read) to locate and parse data files, and uses shell utilities like awk for growth-rate and moving-average calculations.

Analysts and operators who need a data-driven, benchmarked report - not just numbers, but prioritized, actionable recommendations with stated confidence levels, produced according to guidelines that call for objectivity, statistically validated significance, always comparing metrics against baselines, targets, or benchmarks, suggesting charts or graphs for key findings, and documenting methodology.