Data Mapping for Food Insecurity Analysis

Food insecurity, defined as the lack of consistent access to enough food for an active, healthy life, is a pressing issue in the United States. The Food and Agricultural Organization of the United Nations has traditionally measured food insecurity under four dimensions: availability, access, utilization (behavioral choice), and stability (disaster, political instability). In stable environments, hunger and food insecurity are driven by systemic and structural factors that create issues of availability (e.g., food deserts are places with lack of nutritious food supply), affordability (i.e., economic ability to afford nutritious food), and access (e.g., lack of public transportation may act as constraints to accessing a grocery store). While availability causes geographical inequities, affordability and access can additionally cause inequities within a geography. Food insecurity in a geographical region is thus a combination of the three dimensions. Quantifying the contribution of each dimension would be essential to formulating suitable solutions.

However, there are several challenges to quantifying food insecurity and contributing factors. To start, surveys on food insecurity typically do not gather data on contributing features such as public programs (housing, mental health, or improved transportation) while datasets on public programs are in disparate datasets and do not necessarily collect data on food insecurity. Second, surveys on food insecurity are typically aggregated at the state level, but food insecurity can significantly vary across a state. Identifying suitable interventions should be preceded by predictions of food insecurity at more granular county or census-tract levels. In addition, because of synergistic interactions between features, estimates of total food insecurity is not the direct sum of contributions from each dimension; e.g., fewer grocery stores in lowest income neighborhoods,with limited transportation to grocery stores is typical.

Student role: 

The objective of this exploratory research is development of a preliminary data mapping through review of datasets and published literature. Specifically, the student will 

a) Identify appropriate features through literature review: Review literature to identify causal factors and relevant metrics (e.g., transportation metric that can represent accessibility dimension), and features (e.g., feature representative of housing insecurity)  

b) Explore relevant datasets:  Identify relevant datasets that contain food insecurity features; and datasets that contain potential causal features related to each of availability, affordability, and access dimensions. Example data sets: National Health and Nutrition Examination Survey (NHANES), National Health Interview Survey (NHIS), SNAP Policy Database), American Community Survey (ACS), NeilsenIQ retail scanner data etc. 

c) Data mapping: Create a Python script that creates a dataset of features mapped through geographical identifiers, e.g., census tract, county, or state-level. The student can expect to identify datasets by exploring the Science Collaborative for Health disparities and Artificial intelligence bias Reduction (ScHARe), a cloud-based platform that hosts, and enables linking, about 200+ public datasets on surveys and public programs. It also provides the infrastructure for intended data mapping objective.