Data Report — CDC Diabetes Health Indicators

SemMap JSON-LD: dataset.semmap.json · RDFa HTML

Overview

Metric	Value
Dataset	CDC Diabetes Health Indicators
Source	UCI dataset 891
Rows	253,680
Columns	22
Discrete	19
Continuous	3
SemMap	SemMap JSON-LD SemMap HTML
Missingness	Not modeled

Variables and summary

variable	inferred	dist
HighBP	discrete	Reported high blood pressure [1]: 108829 (42.90%)
HighChol	discrete	Reported high cholesterol [1]: 107591 (42.41%)
CholCheck	discrete	Cholesterol check in past five years [1]: 244210 (96.27%)
BMI	continuous	28.3824 ± 6.6087 [12, 24, 27, 31, 98]
Smoker	discrete	At least 100 cigarettes ever [1]: 112423 (44.32%)
Stroke	discrete	Stroke diagnosis [1]: 10292 (4.06%)
HeartDiseaseorAttack	discrete	CHD or MI diagnosis [1]: 23893 (9.42%)
PhysActivity	discrete	Physical activity reported [1]: 191920 (75.65%)
Fruits	discrete	Daily fruit consumption [1]: 160898 (63.43%)
Veggies	discrete	Daily vegetable consumption [1]: 205841 (81.14%)
HvyAlcoholConsump	discrete	Alcohol consumption above heavy threshold [1]: 14256 (5.62%)
AnyHealthcare	discrete	Has health care coverage [1]: 241263 (95.11%)
NoDocbcCost	discrete	Cost prevented doctor visit [1]: 21354 (8.42%)
GenHlth	discrete	Very good health [2]: 89084 (35.12%) Good health [3]: 75646 (29.82%) Excellent health [1]: 45299 (17.86%) Fair health [4]: 31570 (12.44%) Poor health [5]: 12081 (4.76%)
MentHlth	continuous	3.1848 ± 7.4128 [0, 0, 0, 2, 30]
PhysHlth	continuous	4.2421 ± 8.7180 [0, 0, 0, 3, 30]
DiffWalk	discrete	Reported difficulty walking or climbing stairs [1]: 42675 (16.82%)
Sex	discrete	Male [1]: 111706 (44.03%)
Age	discrete	60–64 years [9]: 33244 (13.10%) 65–69 years [10]: 32194 (12.69%) 55–59 years [8]: 30832 (12.15%) 50–54 years [7]: 26314 (10.37%) 70–74 years [11]: 23533 (9.28%) 45–49 years [6]: 19819 (7.81%) 80 years or older [13]: 17363 (6.84%) 40–44 years [5]: 16157 (6.37%) 75–79 years [12]: 15980 (6.30%) 35–39 years [4]: 13823 (5.45%) … (+3 more)
Education	discrete	College 4 years or more (college graduate) [6]: 107325 (42.31%) College 1–3 years (some college or technical school) [5]: 69910 (27.56%) Grade 12 or GED (high school graduate) [4]: 62750 (24.74%) Grades 9–11 (some high school) [3]: 9478 (3.74%) Grades 1–8 (elementary) [2]: 4043 (1.59%) Never attended school or only kindergarten [1]: 174 (0.07%)
Income	discrete	$75,000 or more [8]: 90385 (35.63%) $50,000 to <$75,000 [7]: 43219 (17.04%) $35,000 to <$50,000 [6]: 36470 (14.38%) $25,000 to <$35,000 [5]: 25883 (10.20%) $20,000 to <$25,000 [4]: 20135 (7.94%) $15,000 to <$20,000 [3]: 15994 (6.30%) $10,000 to <$15,000 [2]: 11783 (4.64%) Less than $10,000 [1]: 9811 (3.87%)
Diabetes_binary	discrete	Prediabetes or diabetes diagnosis [1]: 35346 (13.93%)

Fidelity summary

model	backend	disc jsd mean	disc jsd median	cont ks mean	cont w1 mean	downstream sign match
metasyn	metasyn	0.0288	0.0209	0.4691	2.2955	0.5185
clg_mi2	pybnesian	0.0229	0.0186	0.2691	2.9528
semi_mi5	pybnesian	0.0239	0.0157	0.2634	2.9669
ctgan_fast	synthcity	0.2127	0.1539	0.802	7.7665
tvae_quick	synthcity	0.0864	0.0658	0.3607	1.8369

Privacy summary

model	backend	n real	n synth	near duplicate rate eps	nn distance mean	k min	k pct lt5	k map	delta presence
metasyn	metasyn	253680	1000	0.876	0.1199	1	0.9939	1	79
clg_mi2	pybnesian	253680	1000	0.945	0.074	1	0.9939	2	14.5
semi_mi5	pybnesian	253680	1000	0.899	0.0977	1	0.9939	12	1.9333
ctgan_fast	synthcity	253680	256	0.1367	0.3833	1	0.9939	5	3.8
tvae_quick	synthcity	253680	256	0.9258	0.0772	1	0.9939	2	4

Models

UMAP Details Structure

Real data

Model: metasyn (metasyn)

Seed: 42, rows: 1000
Missingness: wrapped (random state 42) via pipeline
Synthetic CSV
Per-variable metrics
Metrics JSON
Privacy metrics
Downstream metrics

Per-variable fidelity

variable	type	KS	W1	JSD
HighBP	discrete			0.0346
HighChol	discrete			0.0356
CholCheck	discrete			0.0302
BMI	continuous	0.084	0.7286
Smoker	discrete			0.0053
Stroke	discrete			0.032
HeartDiseaseorAttack	discrete			0.0198
PhysActivity	discrete			0.0094
Fruits	discrete			0.0212
Veggies	discrete			0.0017

Downstream metrics

metric	value
sign_match_rate	0.5185
formula	Diabetes_binary ~ Q('HighBP') + Q('HighChol') + Q('CholCheck') + Q('BMI') + Q('Smoker') + Q('Stroke') + Q('HeartDiseaseorAttack') + Q('PhysActivity') + Q('Fruits') + Q('Veggies') + Q('HvyAlcoholConsump') + Q('AnyHealthcare') + Q('NoDocbcCost') + Q('GenHlth') + Q('MentHlth') + Q('PhysHlth') + Q('DiffWalk') + Q('Sex') + Q('Age') + Q('Education') + Q('Income') + Q('HighBP'):Q('HighChol') + Q('HighChol'):Q('CholCheck') + Q('CholCheck'):Q('BMI') + Q('BMI'):Q('Smoker') + Q('Smoker'):Q('Stroke')
skipped_reason

Privacy metrics

metric	value
n_real	253680
n_synth	1000
exact_overlap_rate	0
near_duplicate_rate_eps	0.876
nn_distance_mean	0.1199
k_min	1
k_pct_lt5	0.9939
k_map	1
rare_qi_reproduction_rate	0
delta_presence	79

variable	distribution
HighBP	core.multinoulli
HighChol	core.multinoulli
CholCheck	core.multinoulli
BMI	core.lognormal
Smoker	core.multinoulli
Stroke	core.multinoulli
HeartDiseaseorAttack	core.multinoulli
PhysActivity	core.multinoulli
Fruits	core.multinoulli
Veggies	core.multinoulli
HvyAlcoholConsump	core.multinoulli
AnyHealthcare	core.multinoulli
NoDocbcCost	core.multinoulli
GenHlth	core.multinoulli
MentHlth	core.truncated_normal
PhysHlth	core.truncated_normal
DiffWalk	core.multinoulli
Sex	core.multinoulli
Age	core.multinoulli
Education	core.multinoulli
Income	core.multinoulli
Diabetes_binary	core.multinoulli

Model: clg_mi2 (pybnesian)

Seed: 42, rows: 1000
Params: {"max_indegree": 2, "operators": ["arcs"], "score": "bic", "type": "clg"}
Missingness: wrapped (random state 42) via pipeline
Synthetic CSV
Per-variable metrics
Metrics JSON
Privacy metrics

Per-variable fidelity

variable	type	KS	W1	JSD
HighBP	discrete			0.0233
HighChol	discrete			0.0199
CholCheck	discrete			0.0037
BMI	continuous	0.1348	1.2745
Smoker	discrete			0.0169
Stroke	discrete			0.0373
HeartDiseaseorAttack	discrete			0.0229
PhysActivity	discrete			0.0165
Fruits	discrete			0.0186
Veggies	discrete			0.0072

Privacy metrics

metric	value
n_real	253680
n_synth	1000
exact_overlap_rate	0
near_duplicate_rate_eps	0.945
nn_distance_mean	0.074
k_min	1
k_pct_lt5	0.9939
k_map	2
rare_qi_reproduction_rate	0
delta_presence	14.5

Model: semi_mi5 (pybnesian)

Seed: 42, rows: 1000
Params: {"max_indegree": 5, "operators": ["arcs"], "score": "bic", "type": "semiparametric"}
Missingness: wrapped (random state 42) via pipeline
Synthetic CSV
Per-variable metrics
Metrics JSON
Privacy metrics

Per-variable fidelity

variable	type	KS	W1	JSD
HighBP	discrete			0.0095
HighChol	discrete			0.0426
CholCheck	discrete			0.0053
BMI	continuous	0.0958	1.066
Smoker	discrete			0.0044
Stroke	discrete			0.0219
HeartDiseaseorAttack	discrete			0.0229
PhysActivity	discrete			0.0025
Fruits	discrete			0.0139
Veggies	discrete			0.0309

Privacy metrics

metric	value
n_real	253680
n_synth	1000
exact_overlap_rate	0
near_duplicate_rate_eps	0.899
nn_distance_mean	0.0977
k_min	1
k_pct_lt5	0.9939
k_map	12
rare_qi_reproduction_rate	0
delta_presence	1.9333

Model: ctgan_fast (synthcity)

Seed: 42, rows: 256
Params: {"batch_size": 256, "n_iter": 5}
Missingness: wrapped (random state 42) via pipeline
Synthetic CSV
Per-variable metrics
Metrics JSON
Privacy metrics

Per-variable fidelity

variable	type	KS	W1	JSD
HighBP	discrete			0.2
HighChol	discrete			0.1011
CholCheck	discrete			0.0582
BMI	continuous	0.9613	15.7567
Smoker	discrete			0.083
Stroke	discrete			0.0789
HeartDiseaseorAttack	discrete			0.1539
PhysActivity	discrete			0.3231
Fruits	discrete			0.0336
Veggies	discrete			0.0846

Privacy metrics

metric	value
n_real	253680
n_synth	256
exact_overlap_rate	0
near_duplicate_rate_eps	0.1367
nn_distance_mean	0.3833
k_min	1
k_pct_lt5	0.9939
k_map	5
rare_qi_reproduction_rate	0
delta_presence	3.8

Model: tvae_quick (synthcity)

Seed: 42, rows: 256
Params: {"batch_size": 256}
Missingness: wrapped (random state 42) via pipeline
Synthetic CSV
Per-variable metrics
Metrics JSON
Privacy metrics

Per-variable fidelity

variable	type	KS	W1	JSD
HighBP	discrete			0.0605
HighChol	discrete			0.0984
CholCheck	discrete			0.0122
BMI	continuous	0.2469	2.4764
Smoker	discrete			0.0149
Stroke	discrete			0.0529
HeartDiseaseorAttack	discrete			0.0571
PhysActivity	discrete			0.0839
Fruits	discrete			0.0658
Veggies	discrete			0.1223

Privacy metrics

metric	value
n_real	253680
n_synth	256
exact_overlap_rate	0
near_duplicate_rate_eps	0.9258
nn_distance_mean	0.0772
k_min	1
k_pct_lt5	0.9939
k_map	2
rare_qi_reproduction_rate	0
delta_presence	4