Summarise operational data for decision-making

Use Claude Code to analyse operational data in plain language and surface performance insights.

Problem

Farm operations generated daily machinery logs (fuel usage, breakdowns, hours, output) that sat in databases unused. Extracting insights required manual data analysis or custom reporting tools. Managers needed quick answers about machine performance, operator patterns, and maintenance priorities but had no way to query the data conversationally.

Inputs

Machinery logs from Farm Machinery Log app (fuel, hours, output, breakdowns)
Supabase data schema with structured operational data
Geolocation weather API data integrated into Supabase
Standard benchmark data for machinery performance

Claude’s role

Analyst

The pattern

Enable Claude Code to read structured operational data
Accept plain-language queries about machinery performance, patterns, and anomalies
Query relevant data across time periods and machine types
Identify patterns in breakdowns, operator performance, and weather impact
Cross-reference actual performance against standard benchmarks
Generate actionable reports with specific recommendations

Output

Plain-language reports on machinery usage, anomaly analysis, breakdown patterns correlated with operators and conditions, weather impact analysis, training plans for operators, monthly/quarterly/yearly performance summaries, and recommendations for repairs or replacements based on benchmark deviations.

When this works / when it fails

Works well when:

Sufficient historical data exists (months of logs for pattern detection)
Data capture is consistent and reliable
Standard benchmarks are established for machinery performance
Operations have regular patterns that make anomalies detectable

Breaks down when:

Insufficient data volume for meaningful pattern analysis
Inconsistent data quality or gaps in logging
No baseline benchmarks exist for comparison
Operations are too irregular to establish normal performance ranges

Source

Built and used by me