Data Mining

When efficiency ranks before insight

Keeping it short: If you search for information about Bayesian networks, regression techniques, decision trees (CHAID, CART or other), time series or related methods that you want to apply in business, call us - you just found partners for a conversation.

If you have more time: Data Mining is a set of automated or semi-automatized analytical techniques that aim to recognize patterns observed in complex databases and vast amount of data. Traditional statistics are often insufficient to properly control hidden interactions or too complex to compute when applied to millions of records.

The most common applications of Data Mining are related to predicting facts and behaviors. This could be probability of client's resignation (churn) or likelihood of additional sales (cross-, up-selling). A specific subset of data minig techniques is used for classification - to precisely segment your customers.

Overall, the accent is put on precision and efficiency, not necessarily on understanding. One often builds "black boxes" with complex logics that are hard to explain or even understand.

Since the control over the analysis is limited, it is even more critical to:

• supply high quality data ("garbage in, garbage out")
• understand and pre-analyze the data contents.

As analysts, we don't want to limit ourselves to "black boxes". However we know how to build them. If you're done with in-depth analysis and now need precise prediction or effective direct campaigns, have a word with us. We don't get lost even in random forests*!