We are already familiar with various industry use cases for business analytics. We will now take a sneak peek into the future to see what some of the trends are now and will be in business analytics in the coming years. Many innovations are taking place in the analytical space. The following technologies are becoming mainstream: mobile, real time, and pervasive visualization.
Mobile provides reach. Convenience of information access is the most critical factor to establish the habit of fact-based decision making. Information needs to be available when there is a decision to be made. Business intelligence and data discovery applications are going to be accessed more and more from mobile devices as we move forward.
From pushing static data and alerts into mobile devices to fully interactive mobile applications, the key to a successful mobile BI solution is in delivering the practical and tactical information needed to make immediate decisions. The development life cycle for the next generation of BI applications will be in a matter of days instead of weeks or months. This will likely trigger significant change in how IT operates and manages the software development life cycle and conducts change management.
Real-time data for discovery and exploration is also becoming mainstream and a standard requirement.
Historical data is important to run advanced data mining and predictive analytical models to understand customer profiles, demographics, purchasing history, and buying patterns in comparison with other customers in the same clustering. Retailers now have mobile apps that will capture your aisle location in the grocery store and what you have put into your basket to provide real-time recommendations and coupons, based on the batch analytical model as well as the real-time input. The same real-time recommendations apply when you shop online with the input of your recent clicks and browsed items. Your action of whether you redeem that coupon and respond to that promotion will be further captured and stored back to the historic Hadoop cluster to continue to fine-tune the recommendation engine and model.
A real-time streaming event and historical batch model combination gives organizations the most powerful analytical insight and real-time decision.
Lack of ease of use for less technical employees is often quoted as one of the biggest barriers to BI adoption. There has been tremendous demand to put analysis capabilities in the hands of users, not elite analytics or BI experts. Forrester analysts have pointed out the following key characteristics of advanced visualization as it compares to traditional BI: dynamic data, visual querying, linked multidimensional visualization, animation, personalization, and actionable alerts.
The way we process information is guided by how it is presented to us, including attributes such as color, size, texture, density, and movement that activate our visual sensitivity. Figure 1 illustrates how boldface, size, and orientation can easily affect our ability to process information and data.
FIGURE 1. Power of visualization
As you go through each of the squares in Figure 1 clockwise from the top left, you can see how much easier you are able to count the occurrences of the number 8 with simple changes of boldface, orientation, and size. Colin Ware pointed out in his book Information Visualization (Morgan Kaufmann, 2012) that our brain obtains more information through vision than through all of the other senses combined.
Visualization is where data meets design. It is not just about bells and whistles. The main benefit of data visualization lies in its ability to facilitate how our brain processes information and discerns meaning from large amounts of data. We are seeing more vendor solutions that provide preexisting templates, familiar search functions, and guided visual component selection based on the nature of data.
What’s Coming Next
As mobile, real-time, and pervasive visualization continue to mature, we have seen development in some of the following new trends, including natural language processing, cognitive BI, and GeoSpatial revolution. There are a variety of innovative technologies in the BI space, but these three technologies are likely to see quantum leaps and pervasiveness of their usage in the next few years.
Natural Language Processing
Natural language processing (NLP) is at the center of modern software that processes and understands human language. The heart of Web 2.0 is social media and user-generated content. It presents tremendous potential value. Some of the NLP functions are already included in data discovery tools such as Endeca Information Discovery. Examples include word and sentence tokenization, text classification and sentiment analysis, spelling correction, information extraction, parsing, and theme extraction. We’ve also seen technologies like Siri and a slew of voice-prompt software on other mobile devices. The challenges we face stem from the highly ambiguous nature of natural language. More advanced NLP capabilities will be incorporated into BI and data exploration, including algorithms such as n-gram language modeling, Naive Bayes and Maxent classifiers, sequence models like Hidden Markov Models, probabilistic dependency and constituent parsing, and vector-space models of meaning. These algorithms represent the mix of knowledge-engineered, statistical, and machine-learning techniques to disambiguate and respond to natural language input. They will allow our future BI applications to analyze and process complex human language, use data science techniques to automatically generate hypotheses, evaluate a panel of responses based on relevant evidence, discern meaningful information from them, answer our questions through voice prompt, and continue to get smarter based on outcomes with each iteration and interaction. The user interface for BI in the not-so-distant future is going to be more human-to-human with voice and natural language instead of through traditional computer or handheld input devices.
We are already seeing cognitive technology in the BI space for automated information discovery and reporting. Big data systems today are not limited to querying information stored in predefined views, tagged semantic models, or static data models. Data is being accessed via automated indexing for likely correlations and through guided navigation, as is with Endeca Information Discovery.
Human decision making is a complex process. Rational decision making is characterized with the following steps: We collect lots of information, examine a wide variety of alternatives, and then make decisions that maximize the possibility of obtaining the original objective. Situation awareness and mental models play a key role in decision making also. However, we do not make decisions in a manner consistent with this rational model. Nobel Prize winner Herbert Simon has argued that humans possess bounded rationality. We are cognitively limited, such that we can’t possibly be as comprehensive in our information gathering and analysis as assumed. Our cognitive limitations lead to errors in judgment, not because of a lack of intelligence, but simply because we are human. Psychologists describe these systematic mistakes as cognitive biases.
The new generation of BI technology will be more invested in decision-making support on cognitive orientation to overcome these limitations. There are new research projects that are focused on developing related theories, frameworks, and technologies through combining data warehousing, data mining, cognitive psychology, knowledge management, and decision-making theories to facilitate cognitive business intelligence. These projects examine novel concepts, models, algorithms and system architecture such as ontology and experience representation, situation awareness parsing, data warehouse query construction, and guided situation presentation. In time, machines might just make better decisions than humans.
Geographers were dealing with big data way before the term existed. With the explosion of data sets of all types, GeoSpatial data is also omnipresent, spanning from local to global and across themes ranging from natural hazards to energy to water to geology. According to Anthony Robinson, a professor in geography at Penn State, GeoSpatial revolution involves major transformations in how we navigate, share stories, and, ultimately, make decisions.
With the combination of the speed of the Internet, capabilities of satellite technologies, and software such as Google Earth, we are going to see explosive development in how GeoSpatial solutions will change our lives. We are now a location on the map, and we are the center of the context for everything else on the map around us.
One example of an application is the location-based personal assistant as discussed earlier. We are moving away from having to actively search for something to searches telling us what we might be interested in checking out based on location, time of the day, schedules, past activities, demographic information, and general personal preferences. Crowd sourcing is another example that could facilitate real-time reporting of ground conditions in the case of a natural disaster or emergency situation. With the availability of LI-DAR data (LI-DAR stands for Light Detection and Ranging), 3-D terrain models with laser-point accuracy can now be generated for better urban planning such as shadow analysis to determine the impact of a proposed high-rise on a nearby park. The uses and applications are limitless for the commercial, healthcare, and public sectors.
We are continuing to see an evolutionary convergence of search and geospatial technologies from daily personal activities to disaster response and public safety. With mobile technologies, we are becoming individual sensors. Every piece of information we now share with our families, friends, colleagues, and the public has some GeoSpatial tagging on it. We now have access to this entire information ecosystem — it’s a treasure trove that BI solutions will find limitless applications for, and we will see these solutions woven seamlessly into every part of our lives.