Note: This table combines both YP and US Census sources. Table 2. The 25 most numerous amenities in the USA. Table 3. Top 10 YP amenities by region (mean quantity of amenities per zip code by region). Note: regions are defined according to US Census definitions. EISUre be COMPALiits AAMLEMLLES LiL INEW LOPK, CHICURO GM Lifh. Note: This figure compares selected amenities per zip code within LA County, Cook County and the five county boroughs of New York. amenities across the three cities.” The three cities Table 4 compares the 10 most numerous types of YP amenities across the three cities and Figure | compares the average number of other selected amenities across the three cities.” The three cities One of the founding principles of urban sociology is that there is a distinctive form of urban experi- ence (Simmel, 1971), what Louis Wirth called ‘ur- banism as a way of life’ (Wirth, 2004). Major cities often have more in common with one another than they do with smaller cities and rural areas in their geographic regions, and the forms of well-being and quality of life they promote may be similar as well. This idea should not be taken too far: Chicago urbanism, as Edward Glaeser (2011) suggests, likely contains more ‘Midwestern’ qualities than do New York and LA and downtown Houston contains many ‘Texan’ aspects absent in the other three. To highlight how studying amenities illuminates both the commonalities and the differences in the qualitative character of urban experience, we now focus on the three most populous American cit- ies—New York, LA and Chicago.* Urban variation in amenities Figure 2. Scenes dimensions at the neighbourhood level. Note: This figure shows z scores of selected performance scores for four zip codes with the most body art and piercing studios in USA. They show the strength of these scenes dimensions in these places relative to the national average (scored 0). The strong differences in each zip code’s overall scene profile illustrate how no single amenity and no single dimension defines the total scene. IIE FOI IDDIIDE! II IEEE IID IDI IS III III EI II Note: This figure shows z scores of the mean performance scores of all zip codes in LA County, Cook County and the five county boroughs of New York. They show the strength of these scenes dimensions in these cities relative to the national average (scored 0). Figure 4. Regional differences in scenes dimensions. Note: This figure shows regional variations in selected performance scores. Lighter shades indicate counties in lower quintiles (of sel: expression etc.); darker shades indicate higher quintiles. growin. Note: In Figures 5 and 6, bars show the impact via the standardized regression coefficients for scenes that promote self- expression on nine dependent variables (EntP etc.). Whiskers indicate the 95% confidence interval for each coefficient. Thus, if the whiskers do not cross (or touch) the horizontal line at zero, the results are statistically significant. From left to right, dependent variables are log entertainment patents per capita, log other patents per capita, log high-tech patents per capita, log proportional change in total employment (1994-2001), log proportional change in population (1990-2000), proportional change in per capita income (1990-2000), proportional change in gross rent (1990-2000), difference in proportion of college graduates (1990-2000) and difference in proportion of individuals with postgraduate degrees (1990-2000). These are classic least squares multiple regression point estimates. Control variables are summarized in the appendix: population, rent, education, crime, party voting, race, arts industry clusters and the first factor of a principal components analysis on the 15 dimensions of scenes. Subsequent analysis (not reported here) of an expanded model that included some 20 controls revealed substantively similar results. Figure 6. Technology clusters impact local economic growth more strongly in self-expressive scenes. Note: This figure shows the impact via the standardized regression coefficients of technology clusters on two types of patents, job growth, population growth, change in college graduates and change in postgraduates. The full model was repeated across national quintiles of self-expressive scenes. See the note to Figure 5 for more details on the variables in the model. The appendix details how technology clusters are measured.
