01-intro.Rmd

# Introduction {#intro}

Communities provide important resources to the people who live in them. These
resources might include physical and economic resources
--- shared open space, libraries, commercial establishments, etc. --- as well as less identifiable resources
including a sense of membership and other forms of social capital
[@lochner1999]. Indeed, access to these resources is a primary reason why
communities exist [@muth1971], as well as a long-motivating objective in
transportation infrastructure planning [@hansen1959].

Given the importance of these community resources, it is not surprising that 
so much scholarly attention has been paid to examining the spatial and 
socioeconomic variation in access to them [@handy1997; @witten2003]. What
is surprising is the simplistic and arbitrary definition of many
quantitative resource accessibility measures,  in spite of the widespread
availability of geographical information systems (GIS) software [@logan2019] and
an understanding that proximity to a resource is not the only consideration 
in its use [@dong2006]. Individuals do not always shop at the nearest grocery 
store, nor do they necessarily perceive an 11-minute walk to a park as 
meaningfully different from a 9-minute walk. A measure of access that can
incorporate travel impedance by multiple transportation modes alongside
qualitative attributes of the resources in question would provide a better
theoretical comparison to what people experience and observe in their own 
communities. This measure in turn may result in a different understanding 
of which groups have or do not have good access --- and therefore in different
policy interventions to resolve the access gap --- than more traditionally used
measures [@logan2019; @macfarlane2020]. 

In this paper, we develop utility-based access measures to parks, grocery stores, and
libraries in Utah County, Utah. These measures are based in econometric choice
theory relating continuous multimodal travel impedance to attributes of the
resource. The utility preferences are estimated on location-based services data
obtained from a third-party commercial data aggregator. We then use the model
estimates to construct a composite accessibility measure and examine potential
discrepancies between this measure and a more common travel-time buffer 
measurement. 

The paper begins with a discussion of previous findings relating access to 
community resources with social, health, and equity benefits. We then describe
the methodology employed in this research, which makes use of novel third-party
mobile device.  A results section describes both the estimated choice models and
a comparative analysis; the paper closes with a discussion of several 
limitations of the approach as well as associated opportunities for future research.