The future of Business Intelligence (BI) Analytics rests on emerging technologies that will use special hardware appliances that will effectively mine data at tremendous speeds and perform predictive analysis and extrapolate trends using simulation in real-time. Traditional BI products, including dashboards and ad hoc queries use historical data to generate reports, which only tell the story of what happened in the past. In future, predictive analytics, alerts and extrapolation of trends will become common practice through the real-time exploitation of informational patterns hidden in historical data to project what might happen.
Hardware appliances like IBM's Neteeza and Oracle's Exadata provide extreme performance in both OLTP and OLAP models, allowing for very quick deep dives into the data specifically in the latter environment. These appliances are effectively supercomputers that can chew through large volumes of data very quickly.
In addition, the increasing popularity of columnar databases like Sybase IQ, HP Vertica and ParAccel allow data mining software to access a specific type of data or fact based on an attribute or dimension must faster than traditional row-based databases because in a columnar database, each data page is populated with data from only one column.
Tom Davenport and Jeanne Harris in their book "Competing on Analytics" postulate that in the future more automated decisions using "operational business intelligence" will be made by machines rather than relying on humans. It will no longer be acceptable for companies to wait days or weeks to extract data from transactional OLTP systems, load it into data marts and then run historical reports through queries. There will be more prediction and less reporting. There will be greater use of alerts to notify managers immediately when key indicators are at undesirable levels.
Another area of data mining that is almost non-existent toady is the mining of unstructured or textual data. With the ever increasing popularity of social media there is an urgent need to develop efficient parsers to analyze unstructured textual data and transform them into information that can be used in predictive algorithms to project future behavior.
Many of the above changes will hopefully be driven by companies who want to be at the top of their game and increase customer satisfaction. Ultimately the companies with the best analytics will have the most efficient marketing campaigns and promotions. Their customer service will be the envy of their competitors. They will use just-in-time inventory and employ the best people in the industry compensated directly based on their contributions to the company. They will be the leaders who will outperform their competitors and lead us into the future.
Comments