Do your daily business transactions and communication generate a large amount of records? For most companies the answer is yes. These records, such as sales data, supply logs, customer communication logs, and financial records are an expensive investment and an invaluable asset if properly understood and acted upon. Companies willing to share and leverage information across departments and projects stand to gain a considerable advantage over their competition.
At Anderson Analytics, we employ well-developed data mining and text mining techniques to uncover the hidden information within your internal business databases.
Also known as Knowledge-Discovery in Databases (KDD), data mining uses computational techniques from statistics, machine learning and pattern recognition to search large volumes of data for patterns. We use powerful software and techniques such as CHAID, C&RT, PCA/Factor Analysis, Clustering, Kohonen, Neural Nets, GRI, and Natural Language Processing so that all your data can be leveraged for valuable information.
Analyzing Internal Databases
Data mining projects typically involve analyzing the large volume of data our clients already possess, but are unable to make efficient use of. Sales databases, customer databases, employee databases, call center logs, and server logs are all valuable individually but can be much more valuable when linked to each other or to primary survey data. Most clients are excited and surprised at how meaningful and exponentially actionable information leveraged across various sources can be.
Modeling & Simulation
Many data mining projects benefit from modeling and simulation.
Modeling uses mathematical language to describe the behavior of a system. We can model the way your business and market work, the rational behavior of consumers, or model the Marketing P's and their effects on revenue (marketing mix modeling).
Simulation is an imitation of some real state of affairs or process. The act of simulating something generally entails representing certain key characteristics or behaviors of a selected physical or abstract system. Simulation helps us understand a system and make decisions.
Text Mining
As most information (over 80%) is stored as text, Anderson Analytics is at the forefront of this fairly new field. We use state of the art natural language processing (NLP) software which combines linguistics, statistics, and machine learning to code and classify text data. We can then apply data mining and other analytical techniques to uncover the hidden value of this information or link it with other sources.
These techniques can be used for call center customer files, customer suggestions and complaints, survey open ends, focus group or discussion board conversations, blogs, or it can be combined with projective techniques for exponential learning.
Customer complaints and suggestions are not "yes/no" statements; satisfaction is more complex than a 5-point rating scale. Customers think in complete sentences and express their thoughts with individuality and emotion. Text mining takes processing a large amount of unstructured information to the next level. While surveys with rating scales have long been the cornerstone of market research, used alone they often limit the customers' ability to clearly communicate with businesses.
Customers' responses to open-ended questions or to interviews have been treated anecdotally in marketing research in the past. Now, with the availability of text mining computer technology, these verbatim responses can be processed in large volumes in a short period of time.
Our text mining capabilities transcend the dividing line between qualitative and quantitative research methods and can give both more insightful and reliable results, and therefore provides clients with a distinct information advantage.
Hybrid Methodologies
Hybrid Methodology often refers to combining quantitative and qualitative methods in the same study. Because of our state of the art text mining and data mining techniques we go one step beyond simply combining the methodologies. We actually turn traditionally qualitative methods such as projective techniques and psychological/emotive analysis into a quantitative methodology. This allows for statistically valid results and projections of findings onto the general population.
