Median Income Clustering Effect on Usage of Sundae Toppings for Operators
September 1, 2017
By: Jennifer Parra
Fall 2017 GIS 501
Introduction
A food manufacturing and distribution company is attempting to determine the likelihood an operator will use enough volume of sundae topping product to warrant business.  They have decided to focus on this product because of its under-penetration in southern California and Arizona.  A critical component is focusing on a category with the highest level of attractive growth margins, higher levels of innovation, significant opportunities for competitive differentiation and sustainable growth.   Sundae toppings are easily manufactured, have the potential for great diversity, and do not require an increased crop yield for production to increase. They would like to know whether median income clustering has any influence on the volume of product used.  If median income clustering does show a correlation it can give the company information as to whether certain operators are likely to continue to buy product at a rate beneficial to the company.  This information would give the company clear locations for their sales associates to focus business promotion as well.  GIS can be used to develop a clear picture of where there is clustering whether it be low, middle, or high income in southern California and Arizona.  This area is of interest because it is one of the company’s most profitable territories and the two areas have a highly varied socioeconomic structure.  The systems GIS uses to analyze and compile data is well suited for a question of this nature.
GIS as a Tool for Analysis
\citet{Gomarasca_2009} discusses how geomatics is a multidisciplinary and integrated approach to appropriately choosing tools and techniques for displaying spatial data from different domains in a way that is accessible to the user.  Since the question involves several domains into one question and requires multiple steps to formulate conclusions.  So, setting up the parameters, defining thresholds of interest, and determining the order in which to manipulate the data will be important for attaining accurate and valuable results which is reasoning to use GIS to carry out analysis.  Currently regulations are being developed worldwide pertaining to geospatial information and using earth observation data to study and manage phenomena.  We are now able to move towards these globalized initiatives because we now have the capability to house them on the internet in databases that can be accessed by people all over.  This allows the research to become less time consuming, cheaper, and comprehensive research.  This fosters collaboration, participation and the interaction of science with society.
Additionally, \citet{Laurini_2017} discusses how GIS is useful in decision making especially in a specific territory.  "Territorial intelligence can be defined as an approach to regulating a territory ( maybe a city) which is planned and managed by the cross-fertilization of human collective intelligence and artificial intelligence for its sustainable development” \citep{Laurini_2017}.  The information derived from the GIS analysis will give the territory sales representatives a good approach to structuring their strategies for managing different areas.  The information has the ability to give guidance during decision making and planning due to the knowledge that may be derived from the results.  The four main ideas that \citet{Laurini_2017} discusses is using geographic knowledge to explain, manage, monitor, and plan smart territories.  The territories discussed pertain to cities mainly, but the same argument can be made for other territories that have a defined area.  All of the sales managers for this company have very specified territories based on geographic area and population densities.  
Demographics Predicting Use of Funds
One aspect of the proposed question that should be further looked into is the argument that income distribution has any effect on consumers' spending decisions. \citet{Lescaroux2010} proposed this argument in relation to car ownership.  This is a different industry than food service but it can serve as an initial look into how income demographics effect spending habits.  The results indicate that using only the average per capita income does not provide enough information to explain growth patterns and instead the per capita income and the standard deviation of the income distribution could align with past fluctuations in car ownership.  Projections were also able to be made based on the formula that was developed.  This information paired with geographic information could give additional insight into spending decisions.  
\citet{Nayga_1996} analyzed how people were allocating money used for food away from home.  He determined that consumers in the US were spending around 300 billion dollars on the consumption of food away from home.  The company that I am completing this assignment for, manufactures and distributes foods to a wide variety of food away from home (FAFH) operators.  Operators are the companies that dispense the food products to the consumer.  Operators include full service, quick service, fast casual, pubs, clubs, and bars, non-commercial, contract managers, and other distributors.  A consulting firm called L.E.K. conducted a consulting US foodservice operators study in 2014 and  there were 3 main findings of promoting operator business from manufactures and distributors \citep{Piccoiola14}.  Operators need companies to have a better understanding of operators needs, more customized solutions, and more frequent visits from manufacturer reps .  These findings are consistent with ideas discussed in \citet{Nayga_1996}.  In order to keep up with recent trends in consumer demand, a complete understanding of away from home consumption patterns in the US is needed.  The FAFH industry appears to be continuously interested in creating a vast array of concepts that appeal to specific consumer tastes and preferences.  Therefore,  it is imperative that the demographic and socioeconomic profile of individuals be known by the industry that serves them.  
There have been several studies analyzing what types of people spend money on food away from home and convenience foods. \citet{Daniels_2015a} and \citet{Mottaleb_2017} both looked at how different demographic and socioeconomic categories effect usage and habits of food not consumed at home.  The usages were for different countries, Belgium and Bangladesh respectively.  Although the economic structure of these countries may differ from the United States it can still give good insight into common trends of expenditure on food at varying types of restaurants.  \citet{Mottaleb_2017} discusses that food away from home is an established phenomenon in developed countries and that higher income households consume food away from home more frequently than other level households.  This may be due to the monetary ability to do so, the free time that comes with not having to cook, and having no mess to clean up after.  This information will be interesting to test in regard to looking at sundae/dessert topping use distribution.  In the 2015 Daniels article there was not conclusive evidence for consumption of food away from home when compared with age and household composition.  Since preferences can vary drastically between personality types, it would be challenging to parcel out each type of household or personal preference and derive a correlation to product use.  Income level is highly specific and is more objective when determining reasoning for purchasing something or not.   
Conclusions
This social/economic study applies reasoning of a scientific nature to a phenomenon that a company is currently attempting to better understand.  If this question can be answered it has the potential to save the company in operating costs, how much waste is produced, and the time it takes to find new business.  It would allow their sales associates to be more productive in less time and allow them more opportunities to find untapped markets. \citet{Laurini_2017} provides sound reasoning as to why geographic knowledge is beneficial to well functioning territory.  By applying a working hypothesis to the question, it will allow for the method developed to be applied to other territories if needed.  It also leaves several options to look at different demographic or societal factors once the preliminary work has been completed.  Clustering analysis, geocoding, and symbology overlay should provide indications of any correlations between median income clustering and current operator usage.  The software that is currently available through ArcGIS will be sufficient to carry out all analysis if the data is in the correct format prior to the tools being run.  This provides a platform for the data to be stored, analyzed, and a method for distribution.  From the review of literature above it is apparent that median income by itself will likely not be a good predictor of potential sales sites.  There needs to be some additional component that will refine the results. Clustering of median income could be a predictor of product use when overlayed with the current usage at specific operators.