Data mining's value continues to grow for businesses around the globe. In virtually every sector, leading organizations now realize that data mining and analytics technology can be leveraged to achieve powerful results.
The same can also be said of social media. Company leaders understand that Facebook, Twitter, Instagram and other networks are not just invaluable from brand positioning and customer service standpoints – they can also produce invaluable insight into potential and existing clients.
It is therefore unsurprising that an increasing number of organizations are beginning to combine these two resources. By applying data mining and analytics to social media networks, firms can gain a major competitive advantage and even discover new sources of revenue. However, such efforts largely depend on the implementation and maintenance of high-quality data mining tools.
Visual social data
One of the most striking examples of this trend is the rise of image-based social analytics efforts to deliver more targeted marketing. The Wall Street Journal recently reported that a number of different startups are using variations of this approach, and experiencing significant success as a result.
For example, one firm uses visual scanning software to pore over users' photos on Instagram and Facebook in order to identify corporate logos. If the software notices a T-shirt with a company's name on it or a particular product bearing the name of a firm, it also analyzes the broader image, determining whether the featured individuals are smiling and in what context they appear. With this information, companies can identify worthwhile marketing targets, and then carry out more effective campaigns.
The source noted that these efforts are essentially an extension of earlier attempts to data-mine tweets, status posts and other social media text. In those cases, though, analytics success depended upon social media users explicitly referencing and discussing brands and products. Image-based data mining can potentially deliver even more useful insight because photos can offer information about customers' attitudes toward a company indirectly, through people's use of products.
As The Wall Street Journal noted, these tactics are somewhat controversial, as critics worry that such efforts violate users' privacy. However, the source noted that no laws exist preventing organizations from analyzing publicly available images, and all of the photos involved in these data mining projects fit that description.
The amount of available data is tremendous. Instagram alone is home to more than 20 billion users' photos, and tens of millions more are added every day. While not all of these are available to data mining efforts, clearly this is a major resource for marketers.
The source noted that the social media networks hosting these photos allow for visual data mining efforts because they hope companies will consequently choose to advertise on these networks.
The social networks themselves certainly understand the value of data mining for marketing purposes. Notably, Facebook recently announced that it is expanding Audience Network, which mines user data to deliver more targeted ads. Now, Facebook is making this network available to a broader range of marketers around the world, allowing them to move their advertising efforts beyond Facebook itself.
"With the expansion of the Facebook Audience Network, advertisers are getting what they had been hoping for: the ability to use rich social media user data to reach consumers not only on Facebook, but on other mobile apps and services as well," said Debra Aho Williamson, a principal analyst for eMarketer.
Organizations, especially those with no prior data mining experience, are finding it difficult to take the first step towards leveraging analytics. To ease the transition into this and other social-based data mining opportunities, businesses need to embrace proven, high-quality tools, including math and statistical libraries that are easy to embed into existing software. With these tools in place, firms can reduce the cost and time to market for their data mining and analytics projects, and optimize the reliability of the results produced.
• See how to get actionable analytics results faster by reading the IMSL Numerical Libraries datasheet
• In this case study, learn how one company avoided reinventing the wheel and delivered unique analytics results for their customers