PHOTO INTERPRETATION AND HUMAN ECOLOGY OF THE CITY
As described in the following paragraphs and
in the listed references, the photographic
method appears to be particularly useful for
urban studies concerned with population size
and density distributions, and with the pat-
terning of intra-urban socio-economic areas.
The results obtained are contrary to the
prevalent belief that photographic interpre-
tation information in urban area analysis is
limited to physical or material factors.
The research activities summarized herein
had two principal objectives: (1) to test the
accuracy or validity of photo interpretation
data on certain physical-spatial categories by
correlation with observations from ground
surveys, and (2) to develop methods and
techniques for identifying and defining sys-
tematic relationships between the physical-
spatial categories and certain demographic
and social structural characteristics of the
city. Underlying these objectives as a theo-
retical consideration of the urban agglomera-
tion as a complex socio-physical system.
The first task, undertaken in 1952, was the
development and testing of ‘‘keys’’ or criteria
for the identification and classification of
residential structures by aerial photographic
interpretation. The rationale for concentrat-
ing on residences as one of the more impor-
that items of urban physical structure was
based on commonly accepted social values
and sentiment associated with housing and
various types of residential neighborhoods.
One of the key concepts is that a person’s
"address" usually indicates a lot more about
him than just where he lives. Residence loca-
tion has meaning not only in terms of real
estate cost or rental, but also frequently in
terms of occupation, educational level, in-
come class, nationality group, cultural attri-
butes and even religious preference.
As developed and evaluated in the 1952
pilot study of subareas in Birmingham, Ala-
bama, the photo keys proved to be workable
and sufficiently comprehensive, at least in
this first test. The margin of error in identifi-
cation of total numbers of residential struc-
tures was of little or no practical significance.
Some refinement was indicated to improve
the adequacy of the keys for classification of
residences according to family units con-
tained. Also, questions were raised regarding
the application of the method to different
culture areas and geographical regions having
varying architectural features and physical
conditions, all of which would affect its gen-
eral utility. For these reasons, it was appro-
priate to investigate the adaptability of the
2
approach used in the Birmingham study to
another urban complex
Rochester, New York, was selected for this
more intensive follow-up study. This city
contrasts sharply with Birmingham in many
respects. As a result of climatic and sub-cul-
tural variations alone, there are marked dif-
ferences between the two cities in residential
housing. For example, slums in Birmingham
are generally characterized by small, wooden
shanty-like homes crowded together in transi-
tional or interstitial subareas. In Rochester,
on the other hand, most all lower class lodg-
ings are concentrated in large, old converted
structures, outdated apartments, and dilapi-
dated “‘triple-decker’’ tenements. Also, of
course, there are distinct differences in both
the physical settings and economic bases of
the two cities. The combination of all these
contrasts provided the desired conditions for
extending and retesting the methodological
development: (1) an increased variety of
residential structure types, and (2) an in-
creased variety of social, economic, geograph-
ical and cultural characteristics.
The over-all results of the Rochester study
showed a distinct improvement over the find-
ings of the Birmingham pilot study in two
ways. First, for all subareas investigated,
there was an average photo error of only two
per cent in the identification of nearly four
thousand residential structures. Secondly, the
Rochester study was more successful in classi-
fying these structures according to units con-
tained, such as single-unit, duplex," multi-
ple-unit, 3 to 5, 6 to 8, etc. Thus, the follow-
up investigation provided substantial evi-
dence of the adaptability and validity of the
photo keys, the primary focus of interest in
the Rochester undertaking.
In another part of this same study, the in-
vestigators computed several ratios and per-
centages representing various combinations
of structural types existing in the different
subareas of the city. This was in accordance
with the basic objective of developing and in-
terpreting social data from the photographic
information. Subsequently, these structure
ratios were used as predictors in establish-
ing socio-economic status classifications and
rankings of the urban subareas. In addition,
the photo data on dwelling-unit density were
employed in analyses of population density
patterns within the city of Rochester.
As indicated, these latter aspects ofthe
Rochester study were directly related to the
second major phase of the research program,
an empirical investigation of statistical rela-
