78 Gambung green tea samples were analyzed using six gas sensors (TGS822, TGS2602, TGS2620, MQ138, MQ5, MQ3). Each sample was observed three times, generating 60 data records per sensor reading session. The data was labeled for quality standard and organoleptic score by a tea tester referencing the Indonesian National Standard SNI 3945:2016.
Use Cases
- Classify tea quality as 'good' or 'quality defect' based on electronic nose sensor data.
- Predict a continuous organoleptic score based on sensor readings for dry appearance, brew color, taste, aroma, and dregs.
- Train models to correlate specific gas sensor outputs with chemical or physical quality parameters defined by the SNI standard.
- Benchmark electronic nose performance against human sensory evaluation for agricultural products.
Strengths
- Data from 78 distinct tea samples provides a foundation for model training.
- Each sample was observed three times, which may help account for measurement variability.
- Labels are derived from a professional organoleptic test and reference a formal national standard (SNI 3945:2016).
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- The description does not specify the exact sensor features recorded, requiring data inspection.
Provenance
- Source
- Harvard Dataverse, author Dedy Rahman Wijaya.
- Collection Method
- Experimental data collected using a tea chamber and sensor chamber setup, with labeling based on organoleptic testing.
- Freshness
- Last updated 2026-06-01 02:33:34; freshness should be verified.
- Geography
- Likely Indonesia (references Indonesian National Standard), but not explicitly stated.