What data architects can learn from food labels

In software engineering, code refactoring is the process of improving the internal design of source code to make it more readable, reusable and simpler. Refactoring is not just restricted to source code but can also be used on data to improve its information content.

The Food and Drug Administration and the First Lady of the United States (FLOTUS) recently engaged in this kind of “information refactoring” to increase the readability, accuracy and information content of food labels. The proposed changes would increase the font size for the total calorie count per serving, increase the serving size to a more realistic quantity and provide additional information on added sugars.

While the food industry may not agree with all of these changes (and are sure to fight a few of them), it is great to see the FDA and the FLOTUS on the front lines of improving the utility of information provided to consumers. To me, they seem to get the difference between data and information (information being data that is shaped or formed to be useful to its consumers). All government agencies should follow the FDA’s lead and look for opportunities to refactor their data to increase its information content to specific consumers.

In my book, Information As Product, I used a can of food as an analogy to demonstrate the requirements of a good information product. Think about the component parts of a physical product, in this case a can of food, in a grocery store. The total product consists of the food inside the can, the sturdy metal packaging, the food label, the placement on a shelf, the signs on the shelves and aisles, the grocery store building/sign and eventually the advertising that the grocery store chain sends to consumers. Now, let’s map those to our data and data assets to see how we create an information product.

The food itself is the content and content structure (raw data). Everything beyond that is metadata and meta-metadata. The packaging refers to how we organize data into a collection or package of related data. The food label is analogous to key metadata (like the Defense Department’s Discovery Metadata Specification or DDMS) that describes salient characteristics to a general consumer. Of course, the better the producers or marketers know the consumers the more they can tailor a “custom label” to a specific segment of consumers. So, know your consumer! Storing all these “information labels” is akin to the creation of a metadata catalog.

The product aisles and signs are akin to creating contextual metadata that helps explain where the product came from and where it fits into a larger ecosystem (another key technique for this is the creation of taxonomies or categorization schemes). Finally, advertising is a type of push notification to help consumers discover products that may interest them. The information analogue to this are various push notifications, publish and subscribe alerts or email blasts. The point of this is that we can craft effective information products by examining and understanding the construction and marketing of physical products.

So, while data architects can take away valuable lessons from food labeling and marketing, the opposite is also true. How can food labels be improved in the future by leveraging IT innovations like the World Wide Web, linked data and effective metadata techniques? One key lesson from the WWW is that information is not an island; the same rule should apply to food labels. Information should not stop at the physical product. By adding a QR code or somehow mapping existing bar codes to a website, the physical objects can link to virtual information so that a consumer can get more details, corroboration or even critical information about the data on the label.

For example, what if a parent wants more information about a particular food additive that scientists are currently debating? Without a concrete link to a Web page, the chain of information stops and the parent is left to search on his own and form a manual (and perhaps flawed) correlation. Further, that virtual information link could provide additional data about the quality processes, date and place of last inspection to increase consumer trust in the product. Thus our food label could use the same techniques used in the Linked Data community to ensure valuable lineage information extends all the way to the physical product.

Finally, we should have a database of all food labels (or a data set posted to data.gov) so that application developers can create consumer-based applications that exploit that information. The FDA already has a database of drug labels, and the same concept should be implemented for food labels.

Taking a cue from the FDA’s and FLOTUS’s work on food labels, government IT managers should re-examine their data assets and discover new opportunities for “information refactoring.” It’s a low-cost mission multiplier that can improve the productivity of employees, end-users and citizens.

Michael C. Daconta (mdaconta@incadencecorp.com or @mdaconta) is the Vice President of Advanced Technology at InCadence Strategic Solutions and the former Metadata Program Manager for the Homeland Security Department. His new book is entitled, The Great Cloud Migration: Your Roadmap to Cloud Computing, Big Data and Linked Data.