Production Data Analyst

Production Data Analyst

What this skill does

Reads production test reports and time-series production data, computes initial production rates, runs Arps decline curve analysis, checks production allocation, identifies anomalous behavior, and generates production performance summaries. Supports single-well and multi-well pad analysis.

When to use

• User uploads production test reports or production history files
• User asks to analyze production performance or run decline curves
• User wants to compare production against type curve or forecast
• User needs allocation checks across stages or wells
• User is diagnosing liquid loading, anomalous decline, or well interference

Trigger phrases

• "Analyze production data for this well"
• "Run decline curve analysis on this production history"
• "What are the IP30, IP90, and IP180 rates?"
• "Check the production allocation for this pad"
• "Is this well loading up with water?"
• "Compare production against the type curve"
• "Summarize production test results"
• "Flag anomalous production behavior"
• "What is the EUR estimate for this well?"

Inputs

• Production test reports (PDF, DOCX)
• Monthly or daily production time series (CSV, XLSX)
• Pressure and rate data (CSV, XLSX)
• Optional: type curve for benchmarking
• Optional: offset well production for comparison
• Optional: allocation factors per zone or stage

Outputs

• Production test summary with rates and pressures
• IP rates: IP30, IP60, IP90, IP180, IP365
• Peak rate and date of peak
• Decline curve fit parameters (qi, Di, b)
• EUR estimate with confidence range
• GOR and water cut trend
• Allocation table if multi-zone or multi-well
• Anomaly flags (rate spikes, loading indicators, interference signals)
• Structured JSON output

Production metrics computed

Rate metrics

| Metric | Definition | |---|---| | Initial rate (IP) | First production rate or average of first 24 hours | | IP30 | Average daily rate over first 30 producing days | | IP60 | Average daily rate over first 60 producing days | | IP90 | Average daily rate over first 90 producing days | | IP180 | Average daily rate over first 180 producing days | | IP365 | Average daily rate over first 365 producing days | | Peak rate | Maximum recorded daily rate | | Current rate | Most recent 30-day average |

Decline curve analysis (Arps)

| Parameter | Symbol | Description | |---|---|---| | Initial rate | qi | Rate at time zero (bbl/d or MSCF/d) | | Nominal decline rate | Di | Fractional decline per time unit | | Hyperbolic exponent | b | 0 = exponential, 1 = harmonic, 0–1 = hyperbolic | | EUR | EUR | Estimated Ultimate Recovery from decline fit | | Remaining life | — | Years to reach economic limit |

Standard Arps hyperbolic: q(t) = qi / (1 + b × Di × t)^(1/b)

Pressure analysis (when available)

• Tubing head pressure trend
• Casing head pressure trend
• Drawdown = reservoir pressure - flowing BHP
• Buildup analysis if shut-in periods exist

GOR and water cut

| Metric | Definition | Flag condition | |---|---|---| | GOR | Gas-oil ratio (SCF/STB) | Rapid increase may indicate gas breakthrough | | Water cut | Water fraction of liquid production | Rising WC suggests water encroachment | | Liquid loading | Velocity 102% or 3× 30-day average | | LOADING | Rate declining while wellhead pressure increasing | | GORBREAKTHROUGH | GOR increase > 50% over 90 days | | WATERBREAKTHROUGH | Water cut increase > 20 percentage points in 30 days | | INTERFERENCE | Rate decline correlates with offset well startup | | NEGATIVEDECLINE | Rate increasing anomalously without workover | | MISSINGDATA | Gaps > 7 days in production history | | ALLOCATION_ERROR | Allocated sum deviates > 2% from gross |

Output schema

{
  "well_info": {
    "well_name": "string",
    "api": "string",
    "field": "string",
    "first_production_date": "string"
  },
  "ip_rates": {
    "ip30_bopd": 0.0,
    "ip60_bopd": 0.0,
    "ip90_bopd": 0.0,
    "ip180_bopd": 0.0,
    "ip365_bopd": 0.0,
    "peak_rate_bopd": 0.0,
    "peak_date": "string",
    "current_rate_bopd": 0.0
  },
  "decline_curve": {
    "model": "hyperbolic | exponential | harmonic",
    "qi_bopd": 0.0,
    "di_per_year": 0.0,
    "b_factor": 0.0,
    "eur_mbo": 0.0,
    "eur_p10_mbo": 0.0,
    "eur_p90_mbo": 0.0,
    "r_squared": 0.0
  },
  "production_tests": [
    {
      "test_date": "string",
      "oil_rate_bopd": 0.0,
      "gas_rate_mcfd": 0.0,
      "water_rate_bwpd": 0.0,
      "gor_scf_stb": 0.0,
      "water_cut_pct": 0.0,
      "choke_size_in": "string",
      "tubing_pressure_psi": 0.0,
      "casing_pressure_psi": 0.0,
      "test_duration_hrs": 0.0
    }
  ],
  "cumulative_production": {
    "cum_oil_mbo": 0.0,
    "cum_gas_mmscf": 0.0,
    "cum_water_mbw": 0.0,
    "producing_months": 0
  },
  "anomaly_flags": ["string"],
  "quality_flags": ["string"]
}

Decline curve fitting notes

• Require minimum 6 months of production data for reliable Arps fit
• Flag b > 2.0 as physically unlikely (may indicate transient flow still dominating)
• Apply terminal decline rate (typically 8–10% per year) for EUR calculation
• Report P10/P50/P90 EUR range using Monte Carlo or sensitivity on Di ± 20%
• Prefer hyperbolic fit for unconventional wells during boundary-dominated flow

Example response

PRODUCTION PERFORMANCE — Wolfcamp B 3H
════════════════════════════════════════
First production:    June 1, 2024
Producing months:    9
Peak rate:           1,240 BOPD (June 28, 2024)

IP RATES
─────────
IP30:   1,085 BOPD
IP60:     892 BOPD
IP90:     741 BOPD
IP180:    498 BOPD

DECLINE CURVE FIT (Hyperbolic)
──────────────────────────────
qi:   1,240 BOPD
Di:   68% / year
b:    1.4
EUR:  380 MBOE (P50)   |   P10: 490 MBOE   P90: 285 MBOE
R²:   0.94

ANOMALY FLAGS
──────────────
⚠ GOR_BREAKTHROUGH: GOR increased from 850 to 1,420 SCF/STB between months 6–9
⚠ MISSING_DATA: 12-day gap in September (shut-in, reason unknown)

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: ttracx
• Source: ttracx/oil-and-gas-claude-skills
• License: MIT

Install and usage instructions live in the source repository linked above.