The right data and analytic types are valuable assets for business

Business intelligence and data warehousing initiatives involve gathering, manipulating, and analyzing data to turn that data into a more valuable asset and to address a situation or problem. Companies leveraging their information assets are able to rapidly develop and assimilate values, new procedures, and practices into a competitive advantage or more efficient cost structure. And there are numerous data and analytics types that transform business information into valuable assets.

Many corporations only focus on the data that they gather internally. This can be a major oversight since the necessary information or extra value data may only exist from an external source. Data combined from many sources may not have the same or the right context and should be completely analyzed. Data transformation processes need to be consistent, in context, across multiple diverse sources. Data consistency and standard definitions of domains and ranges are vital for developing clear and usable data for decisions and comparison points. Different types of analysis provide different types of comparison points in the warehouse data. The core data context and its related analysis need to be reflective of the type of answers, decisions, and conclusions required by a business. For example, adding weather conditions to the sales figures for various stores can help everyone understand its impact on shopping. For example, with the recent hurricanes in Florida, if the additional weather information were not included, year-to-year sales comparisons would be irrelevant. The business intelligence analysis process should add context, value, and information to help end-users draw conclusions. The business intelligence process should be related to the mission statement and have a definite ordered process. And business intelligence analysis should have a defined process, expectations, and probable actions to be taken. Defining these business intelligence processes helps everyone set their level of expectations and define costs for the analysis outcomes. Defining the analysis, expectations, and action activities helps everyone understand the possible ROI of the complete data warehousing processes. Calculating the ROI can sometimes be very difficult when measuring customer satisfaction or product acceptance or feelings. Focus the data warehousing analytics on the data figures to ?concretely? draw conclusions. Sometimes the assumptions are tested with bias or the wrong data. This is somewhat common because companies only investigate with data from a closed system or from within the company. This is why it may be necessary to include data from outside the company to get a broader point of view. For example, when one school system was recently comparing their costs to others, they found that they were the highest costing district in the state. Comments were made that their budget should be reduced. They then found additional data from other states and compared their costs against other districts around the country with the same size and demographics. Instead, they discovered that their budget was one of the lowest cost centers. If they had not collected the additional data, they would not have known and may have unjustifiably slashed their budget. To really explore and evaluate the full concept of your business intelligence proposition, detail outside conditions and other similar industries that have comparable business procedures. This way you can keep an eye on your end product or action and still explore additional information. All of these business intelligence and data warehouse projects start out with observations, beliefs, or assumptions about a situation. Data is gathered and these assumptions are tested or analyzed to prove or refute the concept. All of these data warehousing and business intelligence analytics can be easily done through several different types of analysis. The most common analytics methods are statistical, association, sequencing, and clustering. Statistical: The first example of analytics is statistical. This form of analysis has been at the core of computing and continues to be extended within database software through various SQL functions like minimum, maximum, standard deviation, regressive average and many others. Leveraging these built-in database analytical functions is much faster than executing them in an outside program since the data does not have to be moved and the operations can take place in a highly tuned memory environment. Association: The next example of business intelligence analysis is through association. Association type analysis mines the data to discover data items that relate to other data items. For example, a family with a median income of $45,000 typically owns a home. This type of analysis takes a given fact and tries to associate other facts with that information. The context of the initial facts and the data to be analyzed should be verified to make sure that the information is in the same context. Having the same context can help ensure that the analysis decisions are based on solid data. The association analysis method takes a known fact and, by analyzing the data, builds other facts. It associates other facts to the base starting point and builds values as more associated trends are identified. For example, companies buy commercials on particular types of television programs to advertise their products. These companies analyze the types of television program, family, car, or news programs. They determine if their product fits with the show. The association between the television show and the product is then extended to an association to the buyer watching that program. Target marketing, at its best, is now defining these associations with more parameters to associate the advertisement flyer to be specific to the individual receiving it. Sequencing: Sequence analysis uses a track record of events to predict or forecast a coming event. This type of analysis usually draws from historical data to determine the future actions or events. This works best by analyzing previous experiences or history of events. It also commonly calculates probabilities and confidence levels of situations and determines their likelihood. For example, sequence analysis is used for determining the sales price of items based on previous sales. Past sales figures are gathered for a similar products selling at various prices. These sale prices and the number of items sold are analyzed to forecast sales of the new item at various prices. By analyzing the different prices and the potential number of items sold, forecasts can be developed to maximize profits. Sequence analysis is also used to forecast sales in a new region. For example, the average sales could be taken from Northern and Southern sales districts and used to project the sales in a new Western region. Sequence leverages previous results and creates enlightened forecast for the future. Pattern analysis is similar to sequence analysis in that it looks for events that correspond to an activity. The difference is that pattern analysis analyzes similarities between data items and groups them together to form patterns. This type of analysis is popular for analyzing buying patterns and frequency or probability of events. Pattern analysis is generally leveraged in retail or telecommunication analysis. Pattern analysis for retail data warehouse applications usually combines the information with lifestyle and life situation data to determine buying and situational patterns. By using pattern analysis, marketers discovered one of the most widely known retail facts, that new mothers tell their husbands to buy baby diapers. By doing pattern analysis on the diaper sales data, they discovered that males were doing the buying and they were also buying beer during the diaper transaction. This discovery led numerous retailers to co-locate several other male type products near the diapers in their stores. Clustering: Cluster analysis evaluates or notes common characteristics and groups them. These groups are then analyzed to determine the desired groups that have the proper characteristics and attributes. Clustering also uses probabilities to determine the likelihood of an additional or one more attribute. These probabilities help determine the potential for the additional characteristics or purchasing situation. Cluster analysis is commonly used on groups of people, demographic or occupation data. For example, statistics have shown that certain occupations are more prone to injuries and the insurance industry uses this type of analysis to determine health and auto insurance rates. Traveling sales people have stressful jobs and driving to their client?s locations adds more risk to their lives, so they have higher insurance rates as a result. Bron: