Food Access Based on Population Density

In Counties of New York State

Project Marvelous-evee
Molly Brakewood, Lauren Chan, Owen Liu, Arthur Sun

5/5/23

Introduce the topic and motivation

  • What is the impact of distance to supermarkets and grocery stores on the food accessibility and food security of low-income households in low population counties of New York?

  • What motivated this was that we wanted to examine the relationship between population size and food accesibility to understand if counties with smaller populations are more susceptible to low food access?

The data we used included information on food accessibility among low-income housing in different counties around the world, but we decided to subset the original data set to include only counties in New York. It was created by Alana Rhone from the Economic Research Service from the CORGIS Dataset Project (2010 census).

Highlights from EDA

  • Research question: Investigating the impact of distance to supermarkets on food accessibility and security for low-income households in low population counties of New York.

  • Data collection and cleaning: Filtering and selecting relevant variables, normalizing low access low-income stats, and creating an analysis-ready data set.

  • Exploratory data analysis: Summarizing mean and standard deviation of low access low-income percentages at different distances.

  • Data limitations: Not having population density information and using data from 2010, which may not reflect current food insecurity trends.

    # A tibble: 1 × 8
  mean_1mile sd_1mile mean_halfmile sd_halfmile mean_10miles sd_10miles
       <dbl>    <dbl>         <dbl>       <dbl>        <dbl>      <dbl>
1       15.6     8.83          21.6        8.89        0.771       1.95
# ℹ 2 more variables: mean_20miles <dbl>, sd_20miles <dbl>

Population Size as an Indicator of Population Density

\[ H_0: r^2 < 0.7 \]

R^2 = 0.78782402536493
We reject H-null. Therefore, we will assume that population size can be used as an indicator of population density for the purposes of this study.

Comparing the Top 10 Largest and Smallest NY State Counties

Hypothesis Testing

Null: There is no difference in the relationship between low access population and distances to the nearest source of food between the 10 largest and 10 smallest counties in New York.

\[H_0: \vert\beta_L - \beta_S\vert = 0\]
Alternative: There is a significant difference in the relationship between low-access population and distances to the nearest source of food between the 10 largest and 10 smallest counties in New York.

\[H_A: \vert\beta_L - \beta_S\vert \neq 0\]

Results:

Observed slope difference: 0.77793 
p-value: 0.774

Using a significance value of 0.05, we can determine our p-value is not statistically significant. Thus, we fail to reject the null hypothesis.

Conclusions + future work

  • Logarithmic relationship proves population size can indicate if a county is rural or urban

  • Low-access population distances to food sources is consistent regardless of county size

  • Highlights the need for targeted interventions and policies to improve access to food in these areas. In the future, we would want to look at data that takes food security initiatives into account (free and reduced school lunch, food banks, etc.) to gain a more in depth understanding of food accessibility in New York State.