The Paradox of Development. In 2008, Google created a system to predict flu outbreaks in advance. Initially, the tool accurately forecasted when and where the flu might occur. However, in 2013, a new feature was added that suggested related search terms. As a result, searchers began inputting their queries less carefully, which led to a decline in the system’s predictive performance. Although improvements have been made since then, this example illustrates how enhancements in one area can cause failures in another. Therefore, it seems like we gain by losing, and often lose by gaining. - Joseph’s “just my thoughts”
Since 2008, the US CDC has published annual flu reports aimed at preventing the spread of influenza nationwide. Researchers took two weeks to compile data on each flu outbreak by calculating the number of cases and generating a report. Meanwhile, the flu had already spread across the country. Google addressed this issue by analyzing search query statistics. However, an error emerged in 2013 when the flu vaccine became scarce, prompting the media to publish numerous flu-related articles that further distorted the situation by conflating the search terms for flu patients. In other words, accurate data analysis depends on the ability to interpret quality, untainted data and its context. - Joseph’s “just my thoughts”