DMA: Data and Marketing Association
Consumer Help

Application of Supervised Machine Learning techniques on Semi-Structured Social Media data for Classification


Twitter feeds, LinkedIn profiles, Pinterest descriptions, and other social media outlets are all comprised of consumer generated text data that is not always clean or neatly lines up with the business problem at hand. The good news is, these data tend to have some structure in the forms of hashtags, character limits, personal tagging, or geo-tagging. This data lends itself to classification schemes developed a priori by the business. This webinar will demonstrate the use of a supervised machine learning technique, Naive Bayes, in order to classify social media posts into these segments. Common data cleansing issues will be discussed as well as methods for evaluating the quality of the results.


Damon Samuel

Director of Data Science
RCG Global Services

Damon Samuel is the Director of Data Science for RCG Global Services and has responsibility of overseeing all advanced analytic applications for the Southeast Region. With nearly 20 years of analytic experience Mr. Samuel is well versed in traditional statistical techniques as well as lead teams on the cutting edge of Big Data applications. Prior to RCG Samuel was responsible for Big Data Analytic Proofs of Concepts as well as Market Mix Modeling at AT&T. Samuel’s work bridges multiple applications from sales forecasting, market mix modeling and attribution, resource optimization, supply chain optimization, fraud detection, risk scoring, loyalty incentive optimization, response modeling and more. He has applied these techniques across telecommunications, staffing, pharmaceuticals, automotive, retail, banking, and hospitality industries. Samuel holds a Masters of Science in Decision Science from Georgia State University and a BA from the University of Missouri.

Sponsored by:

  • DMA Analytics Community. DMA members who sign-up for this webinar will automatically be included in the this community to stay informed of future calls.


Login To Your Account