"An expert in predictive analysis may be able to steer the business toward a better decision that could be made with no analysis at all. Putting your entire analysis budget into software engineers is never a good idea because it could mean that while you have the best analysis software around, nobody knows how to make anything from the data."

Data Analysis for Business

With the reams of data pouring into data warehouses and other storage facilities every second, it can be nearly impossible to make heads of tails of what that data means to you or to your business. It does nobody any good if you have information that would tell you that you need to increase production 30% for the upcoming holiday season or lay off half of your staff because sales are dropping sharply if that information only exists as entries in a database. This is where data analysis comes in.

Data analysis is a process of inspecting the contents of data, cleaning erroneous data out of a data set, transforming that data into a useful format, and putting that data into models that allow you to make decisions. During this process data will be carefully looked over by both people and automated processes to ensure that it is in the correct format, and if not the data will either be discarded or modified until it will fit into existing data structures. After that is completed, the data will be processed, usually by another automated program, into a format that can be loaded into charts, graphs, and other models that allow for analysis of the data. This leaves the information in your database ready to be looked over so decisions can be made or supported.

Data analysis can highlight information such as sales trends, product pricing trends, customer demographics, and even geographical demographics. While it does not take much analysis to make a decision to not ship half of your stock of wool coats to Florida, it may be less clear that while it does not get particularly cold there, leather jackets sell well due to various special events and the fact that the not very cold wind is very humid, and leather insulates well against wet wind.

In addition, data analysis can suggest possible conclusions that may not be apparent at a first glance. Take for example the situation where you need to decide whether or not to open a new factory. You may be aware that you are only producing 70% of the products that are in demand by the public, however data analysis may show you that the demand is seasonal and that opening a new factory may not be the wisest decision. On the other hand, you may use data analysis to examine the productivity of your manufacturing, and find places where you can increase output without the expense of opening an entirely new factory.

There are many disciplines that are all a component of data analysis. Data mining, business intelligence, statistics, exploratory data analysis, predictive analysis, data integration, data visualization, and even data modeling are all aspects of data analysis. Each of these disciplines brings more clarity to the process by either ensuring that the data accurately represents the real world situation, to processing that data to allow for modeling and other analysis, all the way to extrapolating that data to allow for future decision making.

One of the greatest challenges in data analysis and data management in general is trying to decide what data is important enough to keep, and what data is not needed and can be discarded. While analyzing some obscure data sets may lead to surprising results, the additional cost involved can be prohibitive for the relatively small benefits gained by knowing that, for example, persons who purchase a particular brand of pickle also generally purchase a particular brand of cheese. While this information would be useful, especially if you work in the grocery industry and want to know what products to put on sale opposite each other or near each other in the store, that information would take a tremendous amount of additional data storage to obtain.

Almost as important as the analysis itself is the decisions that are made using that data. It is important when building a team to handle analysis that you will need someone who can look at the data and bring forward useful conclusions about that data. Anybody can look at a table containing daily sales volume, but it takes a real expert in data modeling to find out why sales are dropping or increasing, and what could perhaps be done about it.

Predictive analysis allows for the use of data gained through analysis to make decisions about the future. By looking at past trends and what information may be available about the market and competing or complimentary products. An expert in predictive analysis may be able to steer the business toward a better decision that could be made with no analysis at all. Putting your entire analysis budget into software engineers is never a good idea because it could mean that while you have the best analysis software around, nobody knows how to make anything from the data.