CHART 09 / SAMPLE DATASET

Sample Dataset

A synthetic 10-patient sample per department, showing what a merged patient record looks like once registration, the ward, and the lab all feed the same row — and whether that shape of data is workable for a model like XGBoost.

9 Departments Synthetic Data Schema Demo

Where each column comes from

A real patient record here isn't typed in one sitting — it's assembled from three separate points of care. The colour of each column below shows which:

IoT Vitals Kiosk (Height, Weight, BMI, vitals) — captured at registration Haematology / Biochemistry / Pathology — comes back after samples are sent Doctor's desk — entered while using the Clinical Examination Template Provisional Diagnosis — the target column

Browse a department

Patient IDAgeSexFeverCoughBreathlessnessGeneralized weakness / fatigabilityAbdominal painChest painHeadacheLoss of appetite / weight lossNausea / vomitingSwelling of feet (pedal edema)PallorIcterusCyanosisClubbingLymphadenopathyEdemaJVPThyroid swellingHeight (cm)Weight (kg)BMIPulse (bpm)Temp (F)BP (mmHg)SpO2 (%)Resp. Rate (/min)Blood GroupHb (g/dL)TLC (/cumm)ALC (/cumm)Platelet count (/cumm)ESR (mm/hr)CRP (mg/L)RBS (mg/dL)HbA1c (%)Bl. Urea (mg/dL)Sr. Creatinine (mg/dL)Sr. Uric acid (mg/dL)Na (mEq/L)K (mEq/L)Cl (mEq/L)Sr. Bilirubin Total (mg/dL)Sr. Bilirubin Conjugated (mg/dL)Sr. Bilirubin Unconjugated (mg/dL)SGOT / AST (U/L)SGPT / ALT (U/L)GGT (U/L)Albumin (g/dL)Globulin (g/dL)Total Protein (g/dL)ALP (U/L)PT INRT3 (ng/dL)T4 (µg/dL)TSH (µIU/mL)Total Cholesterol (mg/dL)Triglyceride (mg/dL)LDL (mg/dL)HDL (mg/dL)VLDL (mg/dL)MCV (fL)Provisional Diagnosis
MED-00158M1110000100AbsentAbsentAbsentAbsentAbsentAbsentNot raisedAbsent17167.523.1101101.6118/749027B+14.71259224712291463866744.8230.75.51363.61050.730.210.322427184.42.66.9731.21206.53.3214313851352892Community-acquired pneumonia
MED-00231F1110000000AbsentAbsentAbsentAbsentAbsentPresentNot raisedAbsent15656.323.193100.1114/719026O+14871123622326713575.31275.9210.845.614241000.470.080.763925404.32.57.21161.051514.33.32183118102282490Community-acquired pneumonia
MED-00358F0010000001AbsentAbsentAbsentAbsentAbsentAbsentRaisedAbsent16558.821.68998.1127/809122O+14.381462684218408122.8825.6290.94.913041000.560.180.684327334.227861.251168.42.47145176119561996Congestive heart failure
MED-00431F0001000000AbsentAbsentAbsentAbsentAbsentPresentRaisedAbsent14654.425.511397.9115/729122B+118103166927131288.6945.4330.655.51383.61020.880.280.351730334.32.96.7901.0710610.62.9714415676563583Congestive heart failure
MED-00522F0001001000PresentAbsentAbsentPresentAbsentAbsentNot raisedAbsent16164.92510299.3118/749616A+7.5788423712624512041345.4461.034.31364.41001.010.250.592114264.52.96.4881.081028.22.56146144114593459Nutritional anemia
MED-00671M0000000000AbsentAbsentAbsentAbsentPresentAbsentNot recorded / N.A.Absent16266.325.310098.7150/949126A+158977204228415493.3995.5430.53612941040.820.090.463320324.22.58.1980.891247.13.35156125113571686Congestive heart failure
MED-00760F0000001000AbsentAbsentAbsentAbsentPresentPresentNot raisedAbsent15248.921.29498.6109/689922O+15.25721195830701195.31225.1270.34.11383.71060.370.160.634015424.22.47.28911217.52.6120019369691087Tension-type headache
MED-00863M1111000001AbsentAbsentAbsentAbsentAbsentAbsentRaisedAbsent17272.324.49198.3116/739025O+14.169612650263828205.9935.3321.025.21394.7990.960.180.563733384.72.87.3791.081178.13.115521475363593Congestive heart failure
MED-00959M1100000000AbsentAbsentAbsentPresentAbsentAbsentNot raisedAbsent17276.125.795100.4120/758925B+13.31152420392685823739.41125.6230.815.51353.6990.740.230.93525303.82.86.9910.9813010.12.3919215890502689Community-acquired pneumonia
MED-01043M0011000001AbsentAbsentAbsentAbsentAbsentAbsentRaisedAbsent16362.723.69999.1127/809422A+14.98275329521362782.2705.3220.935.61454.6990.530.250.344335253.53.36.91100.9210882.5516615573472791Congestive heart failure

About this data

This is synthetic data built to show the shape a real record would take, not real patient data — values are diagnosis-correlated (not random noise) so the structure is visible, but 10 rows across 4–5 diagnoses per department is far too small to train a model on. It's a schema demonstration: this is what a merged export would look like once real (de-identified) submissions accumulate from the Triage/OPD registration, the vitals kiosk, and the exam templates.

Download the full workbook (.xlsx) — includes a Read Me tab with the same explanation, formulas for BMI, and all nine departments.