Learn by Directing AI
All materials

data-dictionary.md

Data Dictionary

Sensor Readings (sensor-readings.csv)

Hourly water quality readings from IoT sensors installed in 3 of the 8 ponds. Coverage: January 2025 through June 2025 (6 months). Sensors record continuously except during power outages.

Column Type Description
sensor_id String Sensor identifier. Values: SID-001, SID-003, SID-006. Each sensor is installed in one pond.
timestamp Datetime (ISO 8601) Date and time of reading. Hourly frequency. Gaps represent power outages (missing rows, not zero values).
ph Float Water pH. Typical range for vannamei shrimp culture: 7.5-8.5.
dissolved_oxygen_mg_l Float Dissolved oxygen in milligrams per liter. Critical for shrimp survival. Below 4.0 mg/L causes stress; below 3.0 mg/L causes mortality. Typical range: 3.0-9.0 mg/L.
temperature_c Float Water temperature in Celsius. Optimal range for vannamei: 28-32 C. Varies diurnally (higher midday) and seasonally (higher in dry season). Typical range: 26-33 C.
salinity_ppt Float Salinity in parts per thousand. Vannamei tolerate wide range but optimal is 15-25 ppt. Typical range in these ponds: 15-30 ppt.

Notes:

  • Three sensors across 8 ponds. Not all ponds have sensor coverage.
  • Missing rows indicate power outages. Each sensor has 2-3 gap periods of 4-12 hours scattered across the 6 months.
  • Approximately 13,000 readings per sensor when complete (24 hours x ~180 days).

Production Records (production-records.csv)

Per-cycle harvest data for all 8 ponds over 2 years. Two harvest cycles per year (wet season and dry season). Each cycle lasts approximately 90 days from stocking to harvest.

Column Type Description
pond_name String Pond identifier. Values: Pond A through Pond H.
cycle_id String Harvest cycle identifier. Format: C1-YYYY or C2-YYYY where C1 is the first cycle (wet season, roughly Jan-Apr) and C2 is the second cycle (dry season, roughly May-Aug). Values: C1-2023, C2-2023, C1-2024, C2-2024.
cycle_start_date Date Date stocking began for this cycle.
cycle_end_date Date Date of harvest for this cycle.
stocking_density_per_m2 Integer Number of post-larvae stocked per square meter. Range: 80-120. Higher density increases yield potential but also disease risk.
survival_rate_pct Float Percentage of stocked shrimp that survived to harvest. Range: 60-90%. Key performance indicator.
avg_weight_g Float Average individual shrimp weight at harvest in grams. Range: 15-25g. Larger shrimp command higher export prices.
feed_conversion_ratio Float Kilograms of feed per kilogram of shrimp produced. Range: 1.3-1.8. Lower is more efficient.
total_yield_kg Float Total harvest weight in kilograms. Derived from stocking density, pond area, survival rate, and average weight.

Notes:

  • 8 ponds x 4 cycles = 32 rows total.
  • Ponds are identified by name (Pond A-H), not by sensor ID.
  • All 8 ponds have production records. Only 3 ponds have sensor data.