We live in the age of Big Data. So why not analyze and apply this data to prevent inedible food sale, food wastage and enhance food preservation?

This is exactly what inspired University of Washington researchers to explore whether data mining can predict food recalls. Called the Unsafe Food Project, it analyzes data gathered from the FDA, matched with product reviews of various food items on Amazon.com, to predict recalls. The goal has been to identify potentially unsafe food products, and the researchers hope their methods could have further practical applications for investigating illness outbreak in future.

There is a growing concern over how much time it takes between complaints about a food item and the actual time it takes to recall that item. This considerable time lag, which is sometimes even a year, can leave consumers at risk of severe allergic reactions and illness.

Typically, when a compliant surfaces about a food and the reactions it is leading to, it takes time for these to be reported. These reports are then studied by officials, and the illnesses are confirmed as the food testing is done simultaneously. Once the government investigation is completed, the food is then recalled.

Take the recent General Mills recall, for example. The recall happened in batches, and the product was out there as investigations and sales continued in parallel.

The Kansas City plant has been continuing its productions and distribution of flour even as cases of E. coli outbreaks in 21 states were linked to flour made there. The first recall came in May and the second July 1. As the FDA investigation continued, 13 of the 46 confirmed outbreak victims were hospitalized, and one has even developed a severe kidney condition that is often fatal.

The first confirmed case is now being viewed as Dec, 21, 2015, while the most recent one is June 25, with lab tests proving that the E. coli infections in victims have matched strains of the pathogen found in the company's flour. General Mills has had to go back and recall flour between the dates of production Nov. 4, 2015, through Feb. 10, 2016, leading to a total of 45 million pounds.

Clearly, this is a complicated and time-consuming process. The intermediate time can be dangerous for consumers, especially if the food is contaminated, spoiled and not just mislabeled, which can also be dangerous.

This risk is what the UW research project aims to minimize. By using Amazon reviews that consumers post online, data researchers try to identify potentially unsafe food products and predict whether a product will be recalled. In their study, they are especially on the lookout for specific terms like "sick," "mold" and "vomit" among both positive and negative reviews — their research showed even positive or five-star reviews may have problems hidden in them.

By linking the substance of the food product reviews and subsequent food recalls, researchers create correlations and model sensitivity for future studies. In the course of their research, they have also gone on to create an online tool that shows historical data of each product and its reviews.

Following the success of the project, the team is all fired up to explore new avenues in this line of research. They aim to seek cooperation and collaboration from sites like Amazon, Twitter and Yelp, among others, to get information in real time. Emerging technologies and these advanced platforms will help in faster data mining.

A recent report from Stericycle ExpertSolutions says the FDA has recalled 80 times more food units in Q2 2016 than Q1. This may indicate that more accurate testing and stricter regulations are enabling these results.

If these efforts are complemented with the big data analytics that the UW research is producing, there can be comprehensive information to prevent reactions from unsafe food. Data can also be used to enhance production and prevent further fallouts during food processing, production, shelving and packaging and during distribution — all steps that have contamination and spoilage risks.

In the end, Big Data may help lead to not just food safety but also to food security.