Developing a Uniform Coding Structure

REQUEST FOR INFORMATION NOTICE This is a Request for Information (RFI) for market research purposes only and is not a Request for Proposals (RFP). A solicitation is not being issued at this time and this notice shall not be construed as a commitment by the Government to issue a solicitation, nor does it restrict the Government to a particular acquisition approach. This RFI is being published to identify industry capabilities and compare the feasibility of developing a uniform coding structure, and preferably an automated programmable mechanism, for classifying incoming data on a frequent basis from grocery store point of sale (POS) datasets and from existing food purchase or consumption surveys, into a common schema. The resulting schema-data dictionary will allow for the appending of other classifications onto the food items, allowing analysts to aggregate foods by commodity or sub-commodity level, by MyPlate category, or by the foods recommended for frequent -less frequent consumption continuum for example. The North American Industry Classification Systems (NAICS) Code proposed for the requirement is 541611 - Administrative Management and General Management Consulting Services. The size standard for NAICS 541611 is $5M. 1.0 BACKGROUND The Supplemental Nutrition Assistance Program (SNAP) is the United States Department of Agriculture's largest nutrition assistance program, enabling millions of low-income Americans to purchase groceries. With their benefits, households may buy a wide variety of foods for home preparation and consumption, as well as seeds and plants to produce food for the household. Benefits may not be used to buy foods sold hot at the point of sale, alcohol, some energy drinks and products deemed to be supplements, or any non-food items. (For a fuller explanation of benefits see the SNAP program page on foods eligible for purchase http://www.fns.usda.gov/snap/faqs.htm#10.) One of the challenges faced when trying to describe the food purchases of any kind is the sheer number of food items available for purchase. According to the Food Marketing Institute, the average supermarket carried nearly 39,000 items in 2010. Within the 39,000 items the primary ingredient may be the same in many items. For example chicken may be fresh, frozen, or canned; whole or parts; with or without its skin. It can be hard to create useful summary information at the commodity level when it is important sometimes to preserve distinctions in form or package size or other characteristic and sometimes it is not. In order to better understand the impacts of program participation on diet and nutrition, FNS has relied on a variety of nationally representative surveys . Some of these surveys have measured household expenditures on food, whereas others have focused on food consumption at home, either at the household or individual level. Those sources included the National Food Stamp Program Survey (NFSPS), collected in 1996, the National Health and Nutrition Examination Survey (NHANES) data and the diary component of the Consumer Expenditure survey . Other USDA efforts, such as the National Household Food Acquisition and Purchase Survey (FoodAPS) , will examine food purchase behavior via survey. This effort is being undertaken to provide a uniform mechanism for coding and aggregating food purchase and consumption data from a variety of sources to inform analysis of food purchase trends. Purchases may be aggregated at the commodity, sub-commodity or other specified category level. Food products could be a single component, e.g. peas, or could be a composite product e.g. lasagna. 2.0 OBJECTIVES Food retail sales data and food consumption survey data can provide valuable insights about food purchasing trends of households and populations of interest with sales data providing more timely information. Previous research suggests that household electronic supermarket sales data may be a useful surrogate measure of some nutrient intakes of individuals within the household, particularly percentages of energy from saturated and total fat . In the U.S., a majority of grocery stores use electronic product codes to identify and track sales items . Regardless of whether it is sales data or survey data, the data needs to be coded in manner that will allow for the analysis of food purchase trends from the broadest category levels to the more detailed sub-commodity levels to answer policy and research questions. Therefore, this project has three (3) objectives: 1) To develop a food categorization and reference schema or faceted classification system which would allow the FNS to classify information from multiple sources, such as point of sale (POS) store data, food purchase or consumption survey data, and from potential other sources, into common categories useful for analysis. Various levels of aggregation would be needed, for example commodity, sub-commodity, product level etc. Existing database taxonomies (such as the Food and Nutrient Database for Dietary Studies-FNDDS) provide useful examples of multiple levels of product breakdowns. The ultimate goal is to create flexible data base structures to automatically code incoming food purchase data for use in further analysis. 2) The data coding structure should be applicable across point-of-sale and survey data sources to allow analysts to respond to queries about the most recently acquired data or historical data purchases on an on-going basis, of particular food items (like strawberries) to a range of food items (like fresh fruit). This project will not focus on the mechanics of data intake. Nor should it be assumed that the FNS will develop the category definitions to be used, beyond what is accepted in general practice. However, whatever schema is devised should be scalable to be applied in a uniform way to data that is gathered in future. 3) Designing an interface through which analysts can access, manipulate and download new data structures at various levels of aggregation through existing computer software programs (such as Microsoft Access, SPSS, and SAS.) Tools obtained from this effort can be used to help analysts inform policy, and better craft nutrition education and environmental initiatives to enhance the food choices of participants. 3.0 SUBMISSION INFORMATION FNS is seeking information on firms capable of: 1) Creating schema /data dictionaries by which to map existing data, on a continual basis, from store POS datasets and from existing food purchase or consumption surveys, into a common database format, focusing on product level analysis. Knowledge of taxonomies of food classes is a must and understanding of food coding systems is desirable, though nutrient analysis is not. 2) Creating automated ways in which to append other classifications onto the food items, allowing analysts to access foods by commodity or sub-commodity level, by My-Plate category, or by the foods recommended for frequent -less frequent consumption continuum for example. 3) Designing an interface through which analysts can access, manipulate and download new data structures at various levels of aggregation using existing software currently run on USDA FNS systems. Interested firms may submit product descriptions or capability statements that demonstrate its ability to perform all elements of the requirement as described in this notice. The capability statement shall address the firms experience in developing food product food grouping classification systems. All responses must be submitted electronically to Christina.carter-kurant@fns.usda.gov no later than 4PM EST, on Tuesday, April 3, 2012. Any information provided by industry to the Government as a result of this RFI notice is strictly voluntary. Responses will not be returned. No entitlements to payment of direct or indirect costs or charges to the Government will arise as a result of contractor submission of responses, or the Government's use of such information. The information obtained from responses to this notice may be used in the development of an acquisition strategy and future RFP.