When it comes to data, big does not always mean better. Businesses, agencies and websites have become obsessed with the excessive collection of information in this age of big data, even though their efforts do not always yield functional results. A recent scientific article aptly used the term “big data hubris” to describe the mistaken assumption that big data can replace traditional data analysis. Certainly, big data have proven useful and is expanding its reach into consumer applications, but it cannot be used as a solution to every existing problem.

Big data are the large-scale aggregation and analysis of information. Large retailers like Walmart have found value in big data, tracking consumer preferences to determine which products are in demand and when to time sales. Since the growth of data availability fostered by the Internet, new uses of big data have been discovered. UPS has used big data to optimize its delivery routes, dating websites to better match its users and even police departments to gather information on past arrests to calculate when and where crimes are likely to occur again.

Just because big data have succeeded in some areas, however, does not mean it should be applied to every field of research. Tech giant Google used big data to collect information on flu outbreaks. Such analyses are typically made by the Centers for Disease Control and Prevention, which has a large collection of patient data, but Google thought it could better predict future flu outbreaks by tracking the location and frequency of Internet searches for flu symptoms. Unfortunately, as Google has discovered, there is not a direct correlation between people who search for flu symptoms and people who have the flu — healthy individuals are interested in flu symptoms, too. As a result, Google Flu Trends was unsuccessful. In fact, Google reported an inordinately high prevalence of the flu for 100 of the 108 weeks data was tracked.

Google suffers from big data hubris. The Google analysts assumed that big data could replace the traditional data collection methods utilized by the CDC, instead of trying to supplement it. Google Flu Trends would be more accurate if the company had worked in conjunction with the CDC instead of trying to operate independently.

While big data have opened new doors for researchers over the past decade, its blind implementation is unwarranted. Data aggregation is a costly endeavor and does not always pay off. Businesses should proceed with caution when considering the use of big data, making sure to allow its collection to augment, not supplant, traditional research methods.

Adam Fazlibegu is a staff columnist. Adam’s Angle is published every Thursday. Email him at [email protected].