When data mining is in the news, it's often not shown in a particularly positive light. It's portrayed as invasive or depersonalizing, used mostly by advertisers and political campaigns. But these stories overlook the fact that data mining and analytics offer far more than purely commercial value. As a number of recent developments have demonstrated, the technology can play a tremendous, powerful role in the public health sector, as well.
Data mining to fight breast cancer
One of the most noteworthy recent examples of the public health benefits of effective data mining could be found in Chicago. There, Kaiser Health reported, health officials were looking for a way to deliver free mammograms to uninsured women, especially African-Americans. African-American women are much more likely to die from breast cancer, yet the screening rate among this demographic in Chicago was very low.
Chicago public health officials hired a data mining firm to improve the city's outreach to this at-risk group. This firm was formed by many of the analytics professionals who helped run President Obama's successful re-election campaign.
"It's a growing trend that some of the techniques first developed for commercial applications are now spinning off for health applications," explained Jonathan Weiner, professor and director of the Johns Hopkins Center for Population Health IT, the news source reported.
The data mining firm analyzed the Census Bureaus' American Community Survey to develop an algorithm that could accurately predict where at-risk populations without health insurance tended to live. According to Jay Bhatt, chief innovation officer and managed deputy commissioner at the Chicago Health Department, this allowed Chicago health officials to mail fliers to 5,000 women who could most likely benefit from free mammograms but were unlikely to have signed up for screenings, the source reported.
As a result of the success of this project, Chicago health officials plan to expand their use of data mining into other areas.
Mental health insight
Another field benefiting from the use of data mining techniques and technology is mental health. As Psych Central reported, computer scientists at Johns Hopkins have developed a means of leveraging data mining in order to gain valuable mental health insight from Twitter.
"Scientists discovered language patterns associated with a variety of mental health disorders."
According to the source, the researchers used these resources to examine tweets issued by users who publicly acknowledged their mental conditions. This allowed the scientists to discover language patterns associated with a variety of mental health disorders, including post-traumatic stress disorder, bipolar disorder, depression and seasonal affective disorder.
Glen Coppersmith, a senior research scientist with Johns Hopkins, explained that it is far more difficult to obtain quantifiable data concerning mental health issues than it is for physical conditions, due to both the complexity of the underlying causes and the social stigma affiliated with mental health problems. He told the news source that these data mining strategies cannot replace traditional survey methods, but they should be seen as viable options to complement those efforts. He also noted that analyzing social media via data mining is far more affordable and less time-consuming than these older options.
Data mining issues
There are a few lessons that organizations should draw from these two successful data mining efforts. First and foremost, data mining can deliver powerful results when users think creatively about how to apply these techniques. Second, it is often very easy to acquire the raw information needed to conduct effective analytics efforts.
Once your company comes up with a potential data mining project, the question becomes how to deliver actionable information. To this end, we can't stress enough the importance of math and statistical libraries that are easy to embed into your code and provide error codes and documentation that help get the most out of the results. These resources are essential for the development and maintenance of high-performance data mining and analytics applications in every industry.