Skip to main content

Table 4 Regression coefficients obtained via the OLS model in NYC, LA, and Buffalo

From: Deriving neighborhood-level diet and physical activity measurements from anonymized mobile phone location data for enhancing obesity estimation

Variables

NYC

LA

Buffalo

(1) Race and ethnicity

 % Black

2.5460***

1.366***

3.5307***

 % Ame Indi and AK Native

– 0.0550

0.0371

0.0154

 % Asian

– 1.5253***

– 1.5574***

– 0.0841

 % Nati Hawa and Paci Island

– 0.0200

– 0.0263

– 0.0763

 % Hispanic or Latino

1.1223***

1.3375***

– 0.0823

(2) Gender, marital status, and age

 % male

0.0611

0.0269

0.1607

 % married

0.0632

0.0549

– 0.4479

 % age 18–29

– 0.5314***

– 0.7602***

– 1.7860***

 % age 30–39

– 0.1871*

– 0.1909**

0.2533

 % age 40–49

0.0983

– 0.0605

0.1242

 % age 50–59

0.1825**

0.0852

– 0.1283

 % age >  = 60

– 0.7922***

– 0.1804*

– 0.5141

(3) Education

 % < highschool

0.4386***

0.7150***

0.5131

(4) Economic status

 Med income

– 0.8246***

– 0.3493***

– 0.8983*

 % unemployment

0.0858

0.0894*

0.1544

 % below poverty line

1.1520***

1.2396***

1.4296***

 % food stamp/SNAP

0.4041***

0.5628***

0.3141

(5) Housing condition

 Median value units built

– 0.3446***

– 0.1955**

– 0.8326*

 Median year units built

0.2663***

– 0.0964*

0.0784

 % renter-occupied housing units

– 0.8679***

0.2952***

0.1380

(6) Urbanicity

 Population density

– 0.4485***

0.0247

– 0.7433***

(7) Three derived diet and physical activity measurements

 Fast-food restaurant visit frequency (vf)

0.6903***

0.1797***

– 0.1673

 Fitness and sports center vf

– 0.2361**

– 0.1660*

0.0294

 Nature park vf

– 0.0436

– 0.0048

– 0.0037

  1. *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001