CARLA HQ
Add to Chrome
Healthcare & Biomedicine Healthcare & Life Sciences use casesNACE Q86.90Binary ClassificationTarget: classOpenML #1464 ↗

Blood Transfusion Service Center

The Blood Transfusion Service Center dataset contains 748 donor profiles from Hsin-Chu City, Taiwan, designed for binary classification to predict whether a donor will donate blood during a targeted mobile drive campaign based on RFMTC (Recency, Frequency, Monetary, Time) behavioral metrics.

✨ Try in CarlaInstall Chrome Extension ↗100% In-Browser WebGPU • Zero Cloud Upload
748Records (Rows)
4Predictive Features
5 / 0Numeric / Categorical
0.0%Missing Value Ratio
0.24461 - AUC Error
1.05sWebGPU Mean Split
BUSINESS CONTEXT & OBJECTIVEPublic Health & Clinical Blood Banking (Q86.90)

Business Objective: Blood Transfusion Service Center

01Business Context

Regional blood transfusion centers rely on mobile collection buses visiting institutions to maintain adequate blood reserves, facing unpredictable supply fluctuations and donor fatigue.

02Analytical Objective

Predict the likelihood of a registered donor returning to donate blood during an upcoming mobile drive based on historical donation frequency, recency, and tenure.

03Economic & Decision Impact

Accurate propensity scoring optimizes high-intent outreach campaigns and reduces marketing expenditure while preventing blood shortages. False negatives cause missed collection opportunities leading to critical inventory deficits, while false positives waste dispatch and communication budgets on unresponsive candidates.

ML Benchmark Narrative

On this 748-row, 4-feature benchmark, tree-based ensembles like XGBoost and LightGBM typically achieve ROC-AUC scores around 0.74–0.76 due to collinearity between frequency and monetary volume. Zero-shot in-context tabular models like TabICL on Carla HQ match or exceed tuned gradient-boosted baselines out-of-the-box without manual hyperparameter search or feature preprocessing.

Source Origin:UCI Machine Learning Repository / OpenML
Creator:Prof. I-Cheng Yeh, King-Jang Yang, and Tao-Ming Ting (Chung Hua University & Blood Transfusion Service Center, Taiwan) (2008)
License:CC BY 4.0
Reference:Paper / Repo ↗
FEATURE SPECIFICATION & DATA DICTIONARY

Target & Feature Column Definitions

Exhaustive business definitions, measurement units, roles, and target variables across all 5 columns.

ColVariable / Feature NameRoleTypeUnit / ScaleDescription & Business Meaning
AclassDonated Blood in March 2007TARGETNUM—
Binary target indicator representing whether the individual donated blood during the March 2007 mobile campaign (2 = Yes / Donated; 1 = No / Did not donate).
Bv1Months Since Last Donation (Recency)FeatureNUMmonths
Number of months elapsed since the donor's most recent blood donation event.
Cv2Total Number of Donations (Frequency)FeatureNUMdonations
Cumulative total number of blood donations made by the donor across their entire history.
Dv3Total Volume Donated (Monetary)FeatureNUMc.c.
Cumulative volume of blood donated by the individual measured in cubic centimeters (c.c.), where standard donation equals 250 c.c.
Ev4Months Since First Donation (Time)FeatureNUMmonths
Number of months elapsed between the donor's very first recorded donation and the evaluation campaign.
DATASET PREVIEW

Interactive Data Table

Explore rows, feature values, and target labels for Blood Transfusion Service Center.

Download ParquetHF Viewer ↗
Loading dataset...
LOCAL EXECUTION TELEMETRY

TabICLv2 WebGPU Benchmark Results

Complete performance metrics on persisted IID or non-IID evaluation splits, executed 100% locally in the browser sandbox.

Mean Split Latency?Mean wall-clock execution time per persisted evaluation split running 100% locally via WebGPU inside Chrome.Explore Latency Guide →
1.05s60 splits • 8 ensembles/split
Total Evaluation Runtime?Total wall-clock duration across all selected splits, including in-context encoding, WebGPU shader execution, and result aggregation.Explore Split Protocol →
63.18sIncludes warmup & sync
Evaluation Protocol?Precomputed row indices ensure the browser and Python runners evaluate identical out-of-sample observations without leakage.Explore Split Protocol →
20×3 SplitsRepeated IID • 60 total
TabICLv2 Score?Primary task performance score (1 - AUC Error) achieved by TabICLv2 across out-of-sample evaluation splits.Explore Metric Formula →
0.24461 - AUC Error
Evaluation SplitTrain RowsTest Rows
Split Duration?Wall-clock inference time taken for this evaluation split running 100% locally via WebGPU.Learn more →
Accuracy?Proportion of correct test predictions across all classes.Google ML Guide & Details →
ROC-AUC?Area under the ROC curve evaluating ranking capability across all thresholds.Google ML Guide & Details →
F1 Score?Harmonic mean of Precision and Recall, robust against class imbalance.Learn more →
Precision?Proportion of predicted positives that were actually positive.Google ML Guide & Details →
Recall?Proportion of actual positive cases successfully captured.Google ML Guide & Details →
R1 / F14982501.09s76.00%0.72500.36170.50000.2833
R1 / F24992491.06s79.92%0.75310.43180.65520.3220
R1 / F34992491.05s80.72%0.77820.41460.73910.2881
R2 / F14982501.03s78.40%0.73690.42550.58820.3333
R2 / F24992491.04s79.12%0.75840.35000.66670.2373
R2 / F34992491.05s79.92%0.77440.43180.65520.3220
R3 / F14982501.04s78.80%0.71580.43010.60610.3333
R3 / F24992491.04s78.31%0.72930.37210.59260.2712
R3 / F34992491.04s79.92%0.81610.40480.68000.2881
R4 / F14982501.04s78.00%0.74420.36780.59260.2667
R4 / F24992491.05s77.51%0.72580.34880.55560.2542
R4 / F34992491.04s81.53%0.78980.46510.74070.3390
R5 / F14982501.07s77.60%0.75560.31710.59090.2167
R5 / F24992491.05s83.13%0.78140.50000.84000.3559
R5 / F34992491.07s75.90%0.71430.34780.48480.2712
R6 / F14982501.07s80.00%0.78550.47920.63890.3833
R6 / F24992491.04s79.92%0.79900.43180.65520.3220
R6 / F34992491.04s77.51%0.70300.31710.56520.2203
R7 / F14982501.03s80.40%0.74640.44940.68970.3333
R7 / F24992491.06s79.12%0.72970.44680.60000.3559
R7 / F34992491.06s79.92%0.78830.43180.65520.3220
R8 / F14982501.07s80.80%0.78770.44190.73080.3167
R8 / F24992491.04s78.31%0.72470.38640.58620.2881
R8 / F34992491.07s78.71%0.75340.37650.61540.2712
R9 / F14982501.04s79.20%0.75710.40910.64290.3000
R9 / F24992491.06s77.91%0.73650.33730.58330.2373
R9 / F34992491.03s81.53%0.78700.48890.70970.3729
R10 / F14982501.03s78.40%0.76840.34150.63640.2333
R10 / F24992491.09s79.52%0.74850.42700.63330.3220
R10 / F34992491.09s79.12%0.73650.43480.60610.3390
R11 / F14982501.03s82.00%0.77620.50550.74190.3833
R11 / F24992491.09s78.31%0.75400.35710.60000.2542
R11 / F34992491.05s77.91%0.74630.36780.57140.2712
R12 / F14982501.05s80.00%0.73430.44440.66670.3333
R12 / F24992491.06s79.92%0.80180.39020.69570.2712
R12 / F34992491.06s78.31%0.71930.43750.56760.3559
R13 / F14982501.09s78.40%0.73410.34150.63640.2333
R13 / F24992491.07s78.31%0.74880.41300.57580.3220
R13 / F34992491.04s78.71%0.74550.37650.61540.2712
R14 / F14982501.06s77.60%0.76420.37780.56670.2833
R14 / F24992491.06s78.71%0.71050.41760.59380.3220
R14 / F34992491.05s79.52%0.77610.38550.66670.2712
R15 / F14982501.07s78.00%0.70680.36780.59260.2667
R15 / F24992491.05s77.91%0.75600.42110.55560.3390
R15 / F34992491.07s80.72%0.80790.38460.78950.2542
R16 / F14982501.01s77.20%0.74250.27850.57890.1833
R16 / F24992491.06s80.32%0.77600.44940.66670.3390
R16 / F34992491.05s78.31%0.76050.43750.56760.3559
R17 / F14982501.04s81.60%0.74150.46510.76920.3333
R17 / F24992491.07s79.92%0.79140.40480.68000.2881
R17 / F34992491.06s77.51%0.73910.40430.54290.3220
R18 / F14982501.11s76.80%0.72140.48210.51920.4500
R18 / F24992491.07s82.33%0.78590.48840.77780.3559
R18 / F34992491.04s78.31%0.79290.28950.64710.1864
R19 / F14982501.04s78.40%0.79430.32500.65000.2167
R19 / F24992491.03s77.91%0.73510.42110.55560.3390
R19 / F34992491.06s79.12%0.69370.40910.62070.3051
R20 / F14982501.01s79.60%0.77100.46320.62860.3667
R20 / F24992491.03s78.71%0.73010.32910.65000.2203
R20 / F34992491.06s80.32%0.78250.44940.66670.3390
Mean ± Std——1.05s79.10% ± 1.46%0.755 ± 0.0290.404 ± 0.0530.6330.301
UNIFIED BENCHMARK LEADERBOARD

Foundation Models vs Traditional ML Baselines

Side-by-side evaluation of our in-browser Carla engine (WebGPU), open-source tabular foundation models, and OpenML baselines.

RankAlgorithm / ModelModel FamilyRuntime
Predictive Accuracy?Proportion of correct test predictions across cross-validation splits.Google ML Guide & Details →
ROC-AUC?Area under the ROC curve evaluating ranking capability.Learn more →
F-Measure?Harmonic mean of precision and recall.Learn more →
Reference / Repo
#1LMTMachine Learning ModelPython79.68%0.75650.7756OpenML #1663722 ↗
#2Google TabFM v1.0Tabular Foundation ModelPyTorch/Python79.41%0.75550.4187google-research/tabfm ↗
#3EXAONE TabularTabular Foundation ModelPyTorch/Python79.28%0.75200.4197LGAI-Research/EXAONE-Tabular ↗
#4Streaming nanotabiclTabular Foundation ModelPyTorch/Python79.16%0.75540.4064soda-inria/nanotabicl ↗
#5Full TabICLv2 (PyTorch Reference)Tabular Foundation ModelPyTorch/Python79.16%0.75540.4064soda-inria/tabicl ↗
#6MultilayerPerceptronNeural Network (MLP)Python79.14%0.74590.7782OpenML #1811039 ↗
#7ClassificationViaRegression M5PDecision TreePython79.14%0.74720.7647OpenML #1741904 ↗
#8nanotabicl VanillaTabular Foundation ModelPyTorch/Python79.12%0.75540.4052soda-inria/nanotabicl ↗
#9Carla Engine (TabICLv2 WebGPU)Tabular Foundation ModelBrowser79.10%0.75480.4042100% In-Browser
#10AdaBoostM1 DecisionStumpAdaBoostPython78.88%0.73470.7675OpenML #570446 ↗
#11J48Decision TreePython78.88%0.70330.7712OpenML #1688018 ↗
#12TabPFN v3Tabular Foundation ModelPyTorch/Python78.72%0.75380.3645PriorLabs/tabpfn ↗
#13JRipMachine Learning ModelPython78.61%0.64770.7698OpenML #568976 ↗
#14LADTreeDecision TreePython78.61%0.71040.7691OpenML #1666205 ↗
#15KernelLogisticRegression RBFKernelLogistic / Linear ModelPython78.48%0.75910.7433OpenML #1664771 ↗
#16MultiScheme REPTreeDecision TreePython78.48%0.69620.7650OpenML #1672297 ↗
#17Bagging LMTBagging EnsemblePython78.48%0.73780.7619OpenML #1677861 ↗
STATISTICAL DISTRIBUTIONS

Variable Schema & Summary Distributions

Observed numerical ranges, category cardinalities, missing rates, and sample values across 748 rows.

ColVariable NameTypeMissingDistinctSummary Stats / DistributionSample Values
AclassTARGETNUM02[1, 2] μ=1.2 σ=0.4222
Bv1NUM031[0, 74] μ=9.5 σ=8.1201
Cv2NUM033[1, 50] μ=5.5 σ=5.8501316
Dv3NUM033[250, 12500] μ=1378.7 σ=1458.91250032504000
Ev4NUM078[2, 98] μ=34.3 σ=24.4982835
FREQUENTLY ASKED QUESTIONS

Frequently Asked Questions: Blood Transfusion Service Center

Common questions regarding the Blood Transfusion Service Center dataset, machine learning task formulations, and in-browser tabular inference.

What is the Blood Transfusion Service Center dataset used for?

It is an established machine learning benchmark used to predict whether a historical blood donor will donate blood during a specific upcoming mobile drive campaign based on RFMTC (Recency, Frequency, Monetary, Time) behavioral metrics.

Can I run zero-server predictions on this dataset in Google Sheets?

Yes, Carla HQ enables zero-server in-context tabular prediction directly in spreadsheets using TabICL foundation models without exposing sensitive donor records to third-party endpoints.

What machine learning models perform best on the Blood Transfusion Service Center dataset?

Gradient boosted trees (LightGBM, XGBoost, CatBoost) and zero-shot tabular foundation models (TabICL / Carla) achieve leading performance, with ROC-AUC scores typically ranging between 0.74 and 0.77.

LIVE EVALUATION IN GOOGLE SHEETS

Test TabICLv2 on Blood Transfusion Service Center Yourself

Open the pre-loaded Google Sheet and let Carla configure the target and task for local, zero-cloud tabular machine learning.

Step 1

Launch Carla

Click once to open the spreadsheet and Carla side panel together.

Step 2

Review the Setup

Carla selects the dataset target and task from this page automatically.

Step 3

Evaluate & Predict

Run predictions and compute metrics with zero server uploads.

✨ Try Blood Transfusion Service Center in CarlaInstall Chrome Extension ↗

Provenance & Attribution

OpenML Dataset 1464

License: PublicData Source: OpenMLOpenML Page: https://www.openml.org/d/1464

Raw Data Downloads

Download ParquetOpenML Task ↗Google Sheet ↗
All Datasets