Data Evaluation Report

Data Evaluation Report


Report created on: March 15, 2023 21:48:30

Created with SDNIST v2.1.0

Data Description

Deidentified (Deid.) Data:

Label Name Label Value
Team MOSTLY AI
Submission Timestamp 3/10/2023 8:29:44
Algorithm Name MOSTLY AI SD Platform
Variant Label tx2019_synthetic_s1


Property Value
Filename MostlyAI_sd_platform_PaulTiwald
Records 9276
Features 20

Following features has out of bound values, and these are not used in the evaluations.

Dropped Feature Out of Bound Values
PINCP '_RARE_'
NPF 9, 10, 12
INDP '_RARE_'
NOC 9

Target Data:


Property Value
Filename tx2019
Records 9276
Features 24

Evaluated Data Features:

Feature Name Feature Description Feature Type Feature Has 'N' (N/A) values?
PUMA Public use microdata area code object of type string False
AGEP Person's age int64 False
SEX Person's gender int64 False
MSP Marital Status object of type string True
HISP Hispanic origin int64 False
RAC1P Person's Race int64 False
HOUSING_TYPE Housing unit or group quarters int64 False
OWN_RENT Housing unit rented or owned int64 False
DENSITY Population density among residents of each PUMA float64 False
INDP_CAT Industry categories object of type string True
EDU Educational attainment object of type string True
PINCP_DECILE Person's total income in 10-percentile bins object of type string True
POVPIP Income-to-poverty ratio (ex: 250 = 2.5 x poverty line) object of type string True
DVET Veteran service connected disability rating (percentage) object of type string True
DREM Cognitive difficulty object of type string True
DPHY Ambulatory (walking) difficulty object of type string True
DEYE Vision difficulty int32 False
DEAR Hearing difficulty int32 False
WGTP Housing unit sampling weight int64 False
PWGTP Person's sampling weight int64 False

Utility Evaluation

K-Marginal Synopsys:


The k-marginal metric checks how far the shape of the deidentified data distribution has shifted away from the target data distribution. It does this using many 3-dimensional snapshots of the data, averaging the density differences across all snapshots. It was developed by Sergey Pogodin as an efficient scoring mechanism for the NIST Temporal Data Challenges, and can be applied to measure the distance between any two data distributions. A score of 0 means two distributions have zero overlap, while a score of 1000 means the two distributions match identically. More information can be found here.

K-Marginal Score: 920

Sampling Error Comparison:

Here we provide a sampling error baseline: Taking a random subsample of the data also shifts the distribution by introducing sampling error. How does the shift from deidentifying data compare to the shift that would occur from subsampling the target data?

K-Marginal score of the deidentified data closely resembles K-Marginal score of a 30% sub-sample of the target data.

Sub-Sample Size Sub-Sample K-Marginal Score Deidentified Data K-marginal score Absolute Diff. From Deidentified Data K-marginal Score
10% 844 920 76
20% 894 920 26
30% 913 920 7
40% 929 920 9
50% 943 920 23
60% 951 920 31
70% 957 920 37
80% 968 920 48
90% 979 920 59
100% 1000 920 80

K-Marginal Score in Each PUMA:

Different PUMA have different subpopulations and distributions; how much has each PUMA shifted during deidentification?

PUMA Score

48-02516: Tarrant County (Southwest)--Fort Worth (Southwest) & Benbrook Cities

 809

48-02515: Tarrant County (West)--Fort Worth City (West)

 826

48-02510: Tarrant County (North)--North Richland Hills (North) & Keller Cities

 827

48-02507: Tarrant County (East)--Arlington City (West)--South of I-30 & East of Loop I-820

 833

48-02102: Johnson County

 859

48-02101: Ellis County

 863



Univariate Distributions:


Here we provide single feature distribution comparisons ordered to show worst performing features first (based on the L1 norm of density differences).

WGTP: Housing unit sampling weight:


WGTP

PWGTP: Person's sampling weight:


PWGTP

AGEP: Person's age:


AGEP

PINCP_DECILE: Person's total income in 10-percentile bins:


PINCP_DECILE

EDU: Educational attainment:


EDU

INDP_CAT: Industry categories:


Feature Value: N (N/A)
Target Data Counts: 3714
Deidentified Data Counts: 3289

INDP_CAT

POVPIP: Income-to-poverty ratio (ex: 250 = 2.5 x poverty line):


Feature Value: 501 (Not in poverty: income above 5 x poverty line)
Target Data Counts: 3401
Deidentified Data Counts: 3448

POVPIP

MSP: Marital Status:


MSP

SEX: Person's gender:


SEX

DPHY: Ambulatory (walking) difficulty:


Feature Value: 2
Target Data Counts: 8117
Deidentified Data Counts: 8139

DPHY

OWN_RENT: Housing unit rented or owned:


Feature Value: 1
Target Data Counts: 7243
Deidentified Data Counts: 7159

OWN_RENT

HISP: Hispanic origin:


Feature Value: 0
Target Data Counts: 7581
Deidentified Data Counts: 7653

HISP

DREM: Cognitive difficulty:


Feature Value: 2
Target Data Counts: 8300
Deidentified Data Counts: 8291

DREM

DEAR: Hearing difficulty:


Feature Value: 2
Target Data Counts: 8865
Deidentified Data Counts: 8938

DEAR

DEYE: Vision difficulty:


Feature Value: 2
Target Data Counts: 9054
Deidentified Data Counts: 9126

DEYE

HOUSING_TYPE: Housing unit or group quarters:


Feature Value: 1
Target Data Counts: 9050
Deidentified Data Counts: 9035

HOUSING_TYPE

RAC1P: Person's Race:


Feature Value: 1
Target Data Counts: 7831
Deidentified Data Counts: 7861

RAC1P

DVET: Veteran service connected disability rating (percentage):


Feature Value: N (N/A)
Target Data Counts: 9099
Deidentified Data Counts: 9100

DVET

DENSITY: Population density among residents of each PUMA:


DENSITY

PUMA: Public use microdata area code:


PUMA



Correlations:


A key goal of deidentified data is to preserve the feature correlations from the target data, so that analyses performed on the deidentified data provide meaningful insight about the target population. Which correlations are the deidentified data preserving, and which are being altered during deidentification?

Kendall Tau Correlation Coefficient Difference:

This chart shows pairwise correlations using a somewhat different definition of correlation. To what extent do the two different correlation metrics agree or disagree with each other about the quality of the deidentified data?

corr_diff

Pearson Correlation Coefficient Difference:

The Pearson Correlation difference was a popular utility metric during the HLG-MOS Synthetic Data Test Drive. Note that darker highlighting indicates pairs of features whose correlations were not well preserved by the deidentified data.

pearson_corr_diff



Linear Regression:


Linear regression is a fundamental data analysis technique that condenses a multi-dimensional data distribution down to a one dimensional (line) representation. It works by finding the line that sits in the 'middle' of the data, in some sense-- it minimizes the total distance between the points of the data and the line. There are more advanced forms of regression, but here we're focusing on the simplest case-- we fit a simple straight line to the data, getting the slope and y-intercept value of that line.

For this metric we're just looking at data from adults (AGEP > 15) and we're only considering the distribution of the data across two features:
  • EDU: The highest education level this individual has attained, ranging from 1 (elementary school) to 12 (PhD). See Appendix of this report for the full list of code values.
  • PINCP_DECILE: The individual's income decile relative to their PUMA. This helps us account for differences in cost of living across the country. If an individual makes a moderate income but lives in a very low income area, they may have a high value for PINCP_DECILE indicating that they have a high income for their PUMA).

The basic idea is that higher values of EDU should lead to higher values of PINCP_DECILE, and this is broadly true. However, it is known that the relationship between EDU and PINCP_DECILE is different for different demographic subgroups. The heatmaps in the left column below show the density distribution of the true data for each subgroup, normalized by education category (so the density values in each column sum to 1; note that when a cell in the heatmap contains too few people (< 20 ), it is left blank; its not expected that the deidentified data will match the original distribution precisely). The regression line is drawn in red over the heatmap, so you can see the relationship between the target data distribution and its linear regression analysis. In the right column for each subgroup we show how the deidentified data's regression line compares to the target data's regression line, along with a heatmap of the density differences between the two distributions. Redder areas are where the deidentified data has created too many people, bluer areas are where it's created too few people.

We've broken this metric down into demographic subgroups so we can see not only how well the privacy techniques preserve the overall relationship between these features, but also whether they preserve how that overall relationship is built up from the different relationships that hold at each major demographic subgroup. It's important that deidentification techniques preserve these distinct subgroup patterns for analysis.

Total Population:

Target Data:

7594 records, 100.0% of adult (>15) data
Regression: 0.58 slope, 0.79 intercept

Deidentified Data:

7686 records, 100.0% of adult (>15) data
Regression: 0.53 slope, 1.37 intercept
density_plot

White Men:

Target Data:

3128 records, 41.19% of adult (>15) data
Regression: 0.62 slope, 1.48 intercept

Deidentified Data:

3146 records, 40.93% of adult (>15) data
Regression: 0.55 slope, 1.97 intercept
density_plot

White Women:

Target Data:

3351 records, 44.13% of adult (>15) data
Regression: 0.57 slope, 0.19 intercept

Deidentified Data:

3453 records, 44.93% of adult (>15) data
Regression: 0.53 slope, 0.81 intercept
density_plot

Black Men:

Target Data:

236 records, 3.11% of adult (>15) data
Regression: 0.79 slope, -0.96 intercept

Deidentified Data:

213 records, 2.77% of adult (>15) data
Regression: 0.6 slope, 0.87 intercept
density_plot

Black Women:

Target Data:

265 records, 3.49% of adult (>15) data
Regression: 0.68 slope, -0.6 intercept

Deidentified Data:

290 records, 3.77% of adult (>15) data
Regression: 0.58 slope, -0.03 intercept
density_plot

Asian Men:

Target Data:

94 records, 1.24% of adult (>15) data
Regression: 0.63 slope, 0.06 intercept

Deidentified Data:

104 records, 1.35% of adult (>15) data
Regression: 0.51 slope, 1.61 intercept
density_plot

Asian Women:

Target Data:

119 records, 1.57% of adult (>15) data
Regression: 0.33 slope, 1.51 intercept

Deidentified Data:

102 records, 1.33% of adult (>15) data
Regression: 0.65 slope, -0.63 intercept
density_plot

American Indian, Alaskan Native and Native Hawaiians (AIANNH) Men:

Target Data:

19 records, 0.25% of adult (>15) data
Regression: 0.64 slope, 0.42 intercept

Deidentified Data:

17 records, 0.22% of adult (>15) data
Regression: 0.42 slope, 2.95 intercept
density_plot

American Indian, Alaskan Native and Native Hawaiians (AIANNH) Women:

Target Data:

31 records, 0.41% of adult (>15) data
Regression: 0.61 slope, -0.46 intercept

Deidentified Data:

24 records, 0.31% of adult (>15) data
Regression: 0.84 slope, -1.52 intercept
density_plot



Propensity Mean Square Error:


Can a decision tree classifier tell the difference between the target data and the deidentified data? If a classifier is trained to distinguish between the two data sets and it performs poorly on the task, then the deidentified data must not be easy to distinguish from the target data. If the green line matches the blue line, then the deidentified data is high quality. Propensity based metrics have been developed by Joshua Snoke and Gillian Raab and Claire Bowen, all of whom have participated on the NIST Synthetic Data Challenges SME panels.

Score: 0.006

Propensities Distribution:

propensity_distribution



PCA:


This is another approach for visualizing where the distribution of the deidentified data has shifted away from the target data. In this approach, we begin by using Principle Component Analysis to find a way of representing the target data in a lower dimensional space (in 5 dimensions rather than the full 22 dimensions of the original feature space). Descriptions of these new five dimensions (components) are given in the components table; the components will change depending on which target data set you’re using. Five dimensions are better than 22, but we actually want to get down to two dimensions so we can plot the data on simple (x,y) axes– the plots below show the data across each possible pair combination of our five components. You can compare how the shapes change between the target data and the deidentified data, and consider what that might mean in light of the component definitions. This is a relatively new visualization metric that was introduced by the IPUMS International team during the HLG-MOS Synthetic Data Test Drive.

Contribution of Features in Each Principal Component:

Principal Component Features Contribution: feature-name (contribution ratio)
PC-0 PWGTP (0.15),WGTP (0.15),HISP (0.11),OWN_RENT (0.1),RAC1P (0.08)
PC-1 WGTP (0.52),PWGTP (0.5),OWN_RENT (0.33),DPHY (0.21),DREM (0.2)
PC-2 DENSITY (0.64),PUMA (0.64),POVPIP (0.24),HOUSING_TYPE (0.18),RAC1P (0.08)
PC-3 PWGTP (0.44),WGTP (0.41),HOUSING_TYPE (0.33),POVPIP (0.22),AGEP (0.17)
PC-4 HOUSING_TYPE (0.4),DEAR (0.36),DEYE (0.26),POVPIP (0.26),DPHY (0.24)

target
deidentified

PCA Queries:


The queries below explore the PCA metric results in more detail by zooming in on a single component-pair panel and highlighting all individuals that satisfy a given constraint (such as MSP = “N”, individuals who are unmarried because they are children). If the deidentified data preserves the structure and feature correlations of the target data, the highlighted areas should have similar shape.

MSP_N: Children (AGEP < 15):


MSP_N
MSP_N



Inconsistencies:


Summary:

Inconsistency Group Number of Records Inconsistent Percent Records Inconsistent
Age 15 0.2%
Work 0 0.0%
Housing 1 0.0%

Age-Based Inconsistencies:

These inconsistencies deal with the AGE feature; records with age-based inconsistencies might have children who are married, or infants with high school diplomas

child_MSP: Children (< 15) can't be married:

2 violations

Example Record:

AGEP DEAR DENSITY DEYE DPHY DREM DVET EDU HISP HOUSING_TYPE INDP_CAT MSP OWN_RENT PINCP_DECILE POVPIP PUMA PWGTP RAC1P SEX WGTP
13 2 7 2 2 2 N 4 0 1 N 6 1 0 224 48-02507 63 6 1 86

child_PINCP_DECILE: Children (< 15) don't have personal incomes:

2 violations

Example Record:

AGEP DEAR DENSITY DEYE DPHY DREM DVET EDU HISP HOUSING_TYPE INDP_CAT MSP OWN_RENT PINCP_DECILE POVPIP PUMA PWGTP RAC1P SEX WGTP
13 2 7 2 2 2 N 4 0 1 N 6 1 0 224 48-02507 63 6 1 86

child_INDP_CAT: Children (< 15) don't have work industries:

2 violations

Example Record:

AGEP DEAR DENSITY DEYE DPHY DREM DVET EDU HISP HOUSING_TYPE INDP_CAT MSP OWN_RENT PINCP_DECILE POVPIP PUMA PWGTP RAC1P SEX WGTP
14 2 5 2 2 2 N 4 1 1 14 N 1 N 272 48-02516 277 1 1 259

adult_N: Adults ( > 14) must specify values (other than N) for all adult features:

5 violations

Example Record:

AGEP DEAR DENSITY DEYE DPHY DREM DVET EDU HISP HOUSING_TYPE INDP_CAT MSP OWN_RENT PINCP_DECILE POVPIP PUMA PWGTP RAC1P SEX WGTP
17 2 7 2 2 2 N 3 4 1 N N 1 N 305 48-02510 165 1 1 103

toddler_DPHY: Toddlers (< 5) naturally toddle, it's not a physical disability:

6 violations

Example Record:

AGEP DEAR DENSITY DEYE DPHY DREM DVET EDU HISP HOUSING_TYPE INDP_CAT MSP OWN_RENT PINCP_DECILE POVPIP PUMA PWGTP RAC1P SEX WGTP
4 2 2 2 2 2 N 1 0 1 N N 1 N 174 48-02101 131 9 1 92

toddler_DREM: Toddlers (< 5) are naturally forgetful, it's not a cognitive disability:

6 violations

Example Record:

AGEP DEAR DENSITY DEYE DPHY DREM DVET EDU HISP HOUSING_TYPE INDP_CAT MSP OWN_RENT PINCP_DECILE POVPIP PUMA PWGTP RAC1P SEX WGTP
4 2 2 2 2 2 N 1 0 1 N N 1 N 174 48-02101 131 9 1 92

Work-Based Inconsistencies:

These inconsistencies deal with the work and finance features; records with work-based inconsistencies might have high incomes while being in poverty, or have conflicts between their industry code and industry category.

Housing-Based Inconsistencies:

These inconsistencies deal with housing and family features; records with household-based inconsistencies might have more children in the house than the total household size, or be residents of group quarters (such as prison inmates) who are listed as owning their residences.

house_OWN_RENT: Individuals who live in houses must specify if they rent or own:

1 violations

Example Record:

AGEP DEAR DENSITY DEYE DPHY DREM DVET EDU HISP HOUSING_TYPE INDP_CAT MSP OWN_RENT PINCP_DECILE POVPIP PUMA PWGTP RAC1P SEX WGTP
20 2 2 2 2 2 N 7 1 1 12 2 0 2 53 48-02101 194 9 2 90



K-Marginal Score Breakdown:


In the metrics above we’ve considered all of the data together; however we know that algorithms may behave differently on different subgroups in the population. Below we look in more detail at deidentification performance just in the worst performing PUMA, based on k-marginal score.

2 Worst Performing PUMA:

Which are the worst performing PUMA?

PUMA Score

48-02516: Tarrant County (Southwest)--Fort Worth (Southwest) & Benbrook Cities

 809

48-02515: Tarrant County (West)--Fort Worth City (West)

 826

Record Counts in 2 Worst Performing PUMA:

Did the deidentified versions of these PUMA have similar population totals to the target versions?

Dataset Record Counts
Target 2567
Deidentified 2567

Univariate Distribution of Worst Performing Features in 2 Worst Performing PUMA:


Which features are performing the worst in each of these PUMA?

PWGTP: Person's sampling weight:


PWGTP

WGTP: Housing unit sampling weight:


WGTP

AGEP: Person's age:


AGEP

EDU: Educational attainment:


EDU

PINCP_DECILE: Person's total income in 10-percentile bins:


PINCP_DECILE

INDP_CAT: Industry categories:


Feature Value: N (N/A)
Target Data Counts: 1031
Deidentified Data Counts: 910

INDP_CAT

POVPIP: Income-to-poverty ratio (ex: 250 = 2.5 x poverty line):


Feature Value: 501 (Not in poverty: income above 5 x poverty line)
Target Data Counts: 962
Deidentified Data Counts: 974

POVPIP

MSP: Marital Status:


MSP

SEX: Person's gender:


SEX

HISP: Hispanic origin:


Feature Value: 0
Target Data Counts: 2016
Deidentified Data Counts: 2071

HISP

OWN_RENT: Housing unit rented or owned:


Feature Value: 1
Target Data Counts: 1923
Deidentified Data Counts: 1930

OWN_RENT

RAC1P: Person's Race:


Feature Value: 1
Target Data Counts: 2206
Deidentified Data Counts: 2227

RAC1P

DEYE: Vision difficulty:


Feature Value: 2
Target Data Counts: 2507
Deidentified Data Counts: 2532

DEYE

HOUSING_TYPE: Housing unit or group quarters:


Feature Value: 1
Target Data Counts: 2480
Deidentified Data Counts: 2460

HOUSING_TYPE

DVET: Veteran service connected disability rating (percentage):


Feature Value: N (N/A)
Target Data Counts: 2511
Deidentified Data Counts: 2528

DVET

DEAR: Hearing difficulty:


Feature Value: 2
Target Data Counts: 2465
Deidentified Data Counts: 2482

DEAR

DPHY: Ambulatory (walking) difficulty:


Feature Value: 2
Target Data Counts: 2231
Deidentified Data Counts: 2243

DPHY

DREM: Cognitive difficulty:


Feature Value: 2
Target Data Counts: 2282
Deidentified Data Counts: 2288

DREM

DENSITY: Population density among residents of each PUMA:


DENSITY

PUMA: Public use microdata area code:


PUMA

Pearson Correlation Coefficient Difference in 2 Worst Performing PUMA:

How are feature correlations performing in each of these PUMA?

pearson_corr_diff

Privacy Evaluation

Apparent Match Distribution:


Quasi-Identifiers:

These features are used to determine if a deidentified record looks like it might be a real person in the target data.

SEX, EDU, RAC1P, INDP_CAT

Records Matched on Quasi-Identifiers:

Based only on the quasi-identifier features, how many deidentified records uniquely match an individual in the target data? What percentage of the deidentified data has apparent real matches?

67, 0.007% of the deidentified records

Percentage Similarity of the Matched Records:

Considering the set of apparent matches, to what extent are they real matches? This distribution shows edit similarity between apparently matched pairs on how many of the 22 features does the deidentified record have the same value as the real record. If the distribution is centered near 100% that means these deidentified records largely mimic target records and are potentially leaking information about real individuals. If the distribution is centered below 50% that means the deidentified records are very different from the target records, and the apparent matches are not real matches.

apparent_match_distribution

Appendix

Data Dictionary:


PUMA: Public use microdata area code:

PUMA Code Code Description
25-00503 Middlesex County--Waltham City, Lexington, Burlington, Bedford & Lincoln Towns
25-00703 Essex County (East)--Salem, Beverly, Gloucester & Newburyport Cities
25-01000 Peabody City, Danvers, Reading, North Reading & Lynnfield Towns
25-01300 Billerica, Andover, Tewksbury & Wilmington Towns
25-02800 Woburn, Melrose Cities, Saugus, Wakefield & Stoneham Towns
48-02510 Tarrant County (North)--North Richland Hills (North) & Keller Cities
48-02102 Johnson County
48-02101 Ellis County
48-02515 Tarrant County (West)--Fort Worth City (West)
48-02507 Tarrant County (East)--Arlington City (West)--South of I-30 & East of Loop I-820
48-02516 Tarrant County (Southwest)--Fort Worth (Southwest) & Benbrook Cities
01-01301 Birmingham City (West)
06-07502 San Francisco County (North & East)--North Beach & Chinatown
06-08507 Santa Clara County (Southwest)--Cupertino, Saratoga Cities & Los Gatos Town
08-00803 Boulder County (Central)--Boulder City
13-04600 Atlanta Regional Commission--Fulton County (Central)--Atlanta City (Central)
17-03529 Chicago City (South)--South Shore, Hyde Park, Woodlawn, Grand Boulevard & Douglas
17-03531 Chicago City (South)--Auburn Gresham, Roseland, Chatham, Avalon Park & Burnside
19-01700 Des Moines City
24-01004 Montgomery County (South)--Bethesda, Potomac & North Bethesda
26-02702 Washtenaw County (East Central)--Ann Arbor City Area
28-01100 Central Region--Jackson City (East & Central)
29-01901 St. Louis City (North)
30-00600 East Montana (Outside Billings City)
32-00405 Las Vegas City (Southeast)
36-03710 NYC-Bronx Community District 1 & 2--Hunts Point, Longwood & Melrose
36-04010 NYC-Brooklyn Community District 17--East Flatbush, Farragut & Rugby
38-00100 West North Dakota--Minot City
40-00200 Cherokee, Sequoyah & Adair Counties
51-01301 Arlington County (North)
51-51255 Alexandria City

AGEP: Person's age:

AGEP Code Code Description
min 0
max 99

SEX: Person's gender:

SEX Code Code Description
1 Male
2 Female

MSP: Marital Status:

MSP Code Code Description
N N/A (age less than 15 years)
1 Now married, spouse present
2 Now Married, spouse absent
3 Widowed
4 Divorced
5 Separated
6 Never married

HISP: Hispanic origin:

HISP Code Code Description
0 Not Spanish/Hispanic/Latino
1 Mexican
2 Puerto Rican
3 Cuban
4 All other Spanish/Hispanic/Latino

RAC1P: Person's Race:

RAC1P Code Code Description
1 White alone
2 Black or African American alone
3 American Indian alone
4 Alaska Native alone
5 American Indian and Alaska Native tribes specified; or American Indian or Alaska Native, not specified and no other races
6 Asian alone
7 Native Hawaiian and Other Pacific Islander alone
8 Some Other Race alone
9 Two or More Races

NOC: Number of own children in household (unweighted):

NOC Code Code Description
N N/A (GQ/vacant)
0 No own children
min 1
max 19

NPF: Number of persons in family (unweighted):

NPF Code Code Description
N N/A (GQ/vacant/non-family household
min 2
max 20

HOUSING_TYPE: Housing unit or group quarters:

HOUSING_TYPE Code Code Description
1 Housing Unit
2 Institutional Group Quarters
3 Non-institutional Group Quarters

OWN_RENT: Housing unit rented or owned:

OWN_RENT Code Code Description
0 Group quarters
1 Own housing unit
2 Rent housing unit

DENSITY: Population density among residents of each PUMA:

DENSITY Code Code Description
min 16.3
max 52864.7

Density Bin: 2 | Bin Range: (309.67, 475.62]

PUMA DENSITY PUMA NAME
48-02101 357.0 Ellis County
48-02102 450.0 Johnson County

Density Bin: 5 | Bin Range: (1121.99, 1723.27]

PUMA DENSITY PUMA NAME
48-02516 1338.0 Tarrant County (Southwest)--Fort Worth (Southwest) & Benbrook Cities

Density Bin: 6 | Bin Range: (1723.27, 2646.76]

PUMA DENSITY PUMA NAME
48-02515 2134.0 Tarrant County (West)--Fort Worth City (West)

Density Bin: 7 | Bin Range: (2646.76, 4065.16]

PUMA DENSITY PUMA NAME
48-02507 3731.0 Tarrant County (East)--Arlington City (West)--South of I-30 & East of Loop I-820
48-02510 3092.0 Tarrant County (North)--North Richland Hills (North) & Keller Cities

INDP: Industry codes:

See codes in ACS data dictionary. Find codes by searching the string: INDP, in the ACS data dictionary

INDP_CAT: Industry categories:

INDP_CAT Code Code Description
N N/A (less than 16 years old/NILF who last worked more than 5 years ago or never worked)
0 AGR: Agriculture, Forestry, Fishing and Hunting
1 EXT: Mining, Quarrying, and Oil and Gas Extraction
2 UTL: Utilities
3 CON: Construction
4 MFG: Manufacturing
5 WHL: Wholesale Trade
6 RET: Retail Trade
7 TRN: Transportation and Warehousing
8 INF: Information
9 FIN: Finance, Insurance, Real Estate
10 PRF: Professional, Scientific and Technical Services
11 EDU: Educational Services
12 MED: Health Care
13 SCA: Social Assistance
14 ENT: Arts, Entertainment, Accommodation, Food Services and Recreation
15 SRV: Other Services
16 ADM: Government, Public Administration
17 MIL: Military
18 UNEMPLOYED

EDU: Educational attainment:

EDU Code Code Description
N N/A (less than 3 years old)
1 No schooling completed
2 Nursery school, Preschool, or Kindergarten
3 Grade 4 to grade 8
4 Grade 9 to grade 12, no diploma
5 High School diploma
6 GED
7 Some College, no degree
8 Associate degree
9 Bachelors degree
10 Masters degree
11 Professional degree
12 Doctorate degree

PINCP: Person's total income in dollars:

PINCP Code Code Description
N N/A (less than 15 years old)
min -9000
max 1341000

PINCP_DECILE: Person's total income in 10-percentile bins:

PINCP_DECILE Code Code Description
N N/A (less than 15 years old
9 90th percentile
8 80th percentile
7 70th percentile
6 60th percentile
5 50th percentile
4 40th percentile
3 30th percentile
2 20th percentile
1 10th percentile
0 0th percentile

POVPIP: Income-to-poverty ratio (ex: 250 = 2.5 x poverty line):

POVPIP Code Code Description
N N/A
min 0
max 500
501 income above 5 x poverty line

DVET: Veteran service connected disability rating (percentage):

DVET Code Code Description
N N/A (No service-connected disability/never served in military
1 0 percent
2 10 or 20 percent
3 30 or 40 percent
4 50 or 60 percent
5 70, 80, 90 or 100 percent
6 Not reported

DREM: Cognitive difficulty:

DREM Code Code Description
N N/A (Less than 5 years old)
1 Yes
2 No

DPHY: Ambulatory (walking) difficulty:

DPHY Code Code Description
N N/A (Less than 5 years old)
1 Yes
2 No

DEYE: Vision difficulty:

DEYE Code Code Description
1 Yes
2 No

DEAR: Hearing difficulty:

DEAR Code Code Description
1 Yes
2 No

WGTP: Housing unit sampling weight:

See description of weights.

WGTP Code Code Description
0 Group quarters place holder record
min 1
max 9999

PWGTP: Person's sampling weight:

See description of weights.

PWGTP Code Code Description
min 1
max 9999