Using data from online geocoding services for the assessment of environmental obesogenic factors: a feasibility study

Table 3 Results of the field validation

Area	Geocoding service	True positives: N (% positive)^a	False positives: N (% positive)	False negatives: N	Sensitivity^b: %
A	Google Maps	19 (63.33)	11 (36.67)	13	59.38
A	OpenStreetMap	15 (88.24)	2 (11.76)	17	46.88
B	Google Maps	58 (89.23)	7 (10.77)	1	98.31
B	OpenStreetMap	12 (100)	0 (0)	47	20.34
C	Google Maps	144 (71.64)	57 (28.36)	63	69.57
C	OpenStreetMap	41 (87.23)	6 (12.77)	166	19.81
D	Google Maps	22 (64.71)	12 (35.29)	11	66.67
D	OpenStreetMap	21 (80.77)	5 (19.23)	12	63.64

Area A: sparsely populated municipality in the south-west of Bavaria
Area B: area in a medium-sized populated major district town near Munich
Area C: area close to the centre within the densely populated city of Munich
Area D: area with a lower density of amenities within the densely populated city of Munich
^aThe percentage of true positives is the positive predictive value (PPV) [PPV = true positives/(true positives + false positives)]
^bSensitivity = true positives/(true positives + false negatives)

ISSN: 1476-072X