Summary
Many things in life are rightly faced, however of the things that it is worth to have more experience, there are those that arise from mistakes made, because the negative result obtained is known, the causes of its context that influenced the error and the try not to repeat it as much as possible.
This list of errors in the construction of a corporate datawarehouse is one of the first articles, which seek to generate a knowledge base and experience from lessons learned in information technology projects, in this specific case in corporate datawarehouse development projects.
This list was created to serve as a utility to all the people who have some participation in the development of a datawarehouse and will allow to evaluate which is the path they are heading and the opportunities for final success, but the most synthesized advice that can be extracted from it is to maintain joint work at the participating levels and organizational areas as well as to work on solid theoretical-practical bases of the technology involved.
The objective of this list is to provide a source of consultation of lessons learned, from the application of this technology in various companies, as well as to provide theoretical-practical basis, which rests on the domain of computer science and more properly, of the datawarehouse technology, and which should ultimately lead to the future success of your company's corporate datawarehouse project.
Keywords: datawarehouse, datamart, datamining, construction of a corporate datahouse, common mistakes, managing a datawarehouse project, risks of building a datawarehouse, prospects of the datawarehouse, profitability metric on the investment of a datahouse
Index
1. Introduction
2. List of 10 Most Common Errors in the Construction of a Corporate Datawarehouse
3. Conclusions
4. Bibliography
1. Introduction
Today, technology areas of most financial and commercial companies, which until recently devoted most of their efforts to providing transactional information systems - that supported the burden of most activities in their value chains - They are focused on achieving the coherent exploitation of their stored data: historical and transactional, the reality around it, is the enormous size of data obtained from the daily operations of their transactional systems, and the problem of analyzing and extracting knowledge from all this. information that remains buried within itself.
There are a series of computational techniques that allow the transformation of this transactional, operational and daily information into information with different levels of aggregation, summarized, exact and specialized by topic, and which allows their analysis, facilitating managerial decision-making, This technology is called a Data Repository or Datawarehouse.
However, and based on professional practice, the construction of a datawarehouse is not immune to problems that impede its success and the achievement of its final objective: support for managerial decision-making.
This article attempts to go beyond the theoretical, to present problematic issues that prevent the successful development of a corporate datawarehouse.
Therefore, it should be specified that the following list - of the most common errors in the development of a corporate datawarehouse - emerges from the development of this technology in representative companies belonging to the Peruvian financial and commercial sector.
Assuming that a project must satisfy the client by representing returns on their investment - we know that it is not always easy to estimate this in computer technology projects, however - the success of a datawarehouse development project is directly proportional to the utility obtained from the analysis of the information in making the right decisions in which the company benefits, at this point: knowing that a decision made - based on the analysis of data through the datawarehouse - produces greater income and / or savings, is the most Adequate estimation, for this we must consider the ease of obtaining reports as well as the flexibility of its configuration.It is considered - to generate learning from the decisions made each time - the importance of storing the form and information of the aggregates and reports used over time, as a way of demonstrating their value for the company.
This list, with proper use, will allow for more accurate decision-making by project management, owners and clients, knowing the possible consequences of one or another determined strategy or measure and preventing frequent errors, allowing cost savings, both in the formulation, development and implementation of the project.
2. List of 10 Most Common Errors in the Construction of a Corporate Datawarehouse
1st Error: Assume the solution of the problems that may arise as a purely technical question
The datawarehouse requires the active participation of managerial decision makers.
2nd Error: Not assigning an adequate budget for the entire project.
An adequate allocation of capital and resources to support and function the technological platform and infrastructure that a corporate datawarehouse requires must be one of the initial aspects when considering it as a project.
3rd Error: Lack of commitment from senior management.
The success of the corporate datawarehouse requires the full support of top management based on the security and trust provided to the project management and its development team, to allow the work to be fluid in all the organizational areas involved in the project.
4th Error: Not having an adequate infrastructure to support it.
The corporate datawarehouse requires an adequate technological as well as organizational infrastructure. The systems architecture of a datawarehouse ranges from proprietary database servers, data transformation and cleaning servers, front-end managerial user nodes arranged throughout the organization. Software requires server, client, and business layer applications that work efficiently in n-dimensional query environments and in parallel processing.
5th Error: Redundant, non-transparent and undocumented databases.
The state of the transactional databases from which the information that will be transformed and stored in the datawarehouse is extracted, is generally not considered in the initial time estimates, however, this may represent significant delay, the ideal in any case is that the states of the entities such as accounts, clients, debts, payments, etc. are stored in historical tables for significant periods of time according to the average frequency of variation (monthly, biweekly, daily) and in standard form, however in practice this information is found as part of tables used for management reports, which have already gone through a filtering process, thus losing the coherence of historical data and changes in their states over time at the level of everything the business.
In some cases it is recommended to do an independent reengineering of databases at a transactional level, in which changes in the states of the entities can be stored in a clean way as a previous step to consider the extraction of the data from its origins.
6th Error: Failure to foster an environment of complete cooperation between DBA's and the datawarehouse team.
When starting a datawarehouse project in a company, the DBAs Database Administrators area generally already exists, so it is recommended to create the Datawarehouse Area at the same level as the DBAs - and not under the control of the DBAs. Many DBAs are responsible for maintaining the databases to support day-to-day transactions. Creating an alternative platform sufficient for the development of the datawarehouse, having access to the information sources of the datawarehouse domain directly and not as intermediaries to the DBAs, is one of the important factors that contributes to a faster development of the project and that aim for success.
7th Error: Failure to use a requirements specification methodology suitable for management.
In the requirements specification activities by management users, not using a methodology that allows the user to express their requirements easily and ready for subsequent feedback is one of the frequent errors.
The use of prototyping is recommended, as well as the preparation of functional specification documents for requirements by management areas.
8th Error: Ignorance of the value chain, information flows in business activities.
Identifying the activities that are critical success factors, as well as observing the flow of activities of central competence that allow the products or services offered to provide value to the client, is a task of knowledge of the business that the designers of the datawarehouse must not pass bypassing this for the choice of a project development strategy - based on a general model of developing datamart by processes or organizational areas - that allows establishing priorities according to the company's strategic plan.
9th Error: Not having an integration perspective with other related technologies.
A solid theoretical base in information technology, and the wide range of existing trends and solutions, will allow the development of a datawarehouse with a vision for the future. In this perspective, related technologies are: OLAP that allows the analysis of historical information to determine patterns of behavior, on the other hand, datamining that allows the discovery of patterns of behavior but automatically using models and algorithms (decision trees, clusters, network neurals, fuzzy logic, linear regression, etc.), Somewhat further away is the implementation of a corporate balanced scorecard, where the datawarehouse serves information from the indicators for historical periods of time.
10th Error: Bad project management and discontinuation of the project development plan.
The capacity of the project manager is exposed in all its real dimension, both in technical and human training and the product of experiences in similar projects. The development of the datawarehouse must be cyclical and in stages, with the activities of analysis, design, development, testing; repetitive milestones, trying to avoid repetitive work.
3. Conclusions
This article is part of a series of articles that the author presents on information technology, particularly in this article, giving a list of 10 of the most common errors that occur in the development of projects to build a corporate datawarehouse..
The theoretical basis on which this class of projects are based starts from the conception of Multidimensional Information Databases (Ontology, Conceptual Data Model, Semantic Interpretation, Theory of Sets and Relations), from which it is expected to obtain exact information., accurate and summarized for analysis and support for managerial decision making.
The practical basis was obtained as a result of the author's participation in various datawarehouse projects in important Peruvian commercial and financial companies, whose experience is impregnated in this article, and is presented to the reader as a tool for consultation and theoretical-practical basis., for future success in developing your corporate datawarehouse.
4. Bibliography
1. Building the data warehouse Author WH Inmon Editorial John
Wiley & Sons, Inc. New York, NY, USA 1996
2. The Data Warehouse Lifecycle Toolkit: Expert Methods for Designing, Developing and Deploying Data Warehouses Authors Ralph Kimball, Laura Reeves, Warren Thornthwaite, Margy Ross, Warren Thornwaite Editorial John Wiley & Sons, Inc. New York, NY, USA 1998
3. Genetic Algorithm for Materialized View Selection in Data
Warehouse Environments Editorial Springer Berlin / Heidelberg
1999
4. Data Warehouse Modeling and Quality Issues Author: Panos Vassiliadis - Knowledge and Data Base Systems Laboratory Computer Science Division - Department of Electrical and Computer Engineering - National Technical University of Athens- Zographou 157 73, Athens, GREECE phd.pdf
5. Maintenance of Data Warehouse Views Using Standardization
Authors: Mukesh Mohania, Kamal Karlapalem, Millist Vincent In D.
Ram, editor, Data Management, pages 32–50. Springer Verlag, 1997.
6. A methodology for datawarehouse design: conceptual modeling Authors José María Cavero Universidad Rey Juan Carlos, Spain, Esperanza Marcos Universidad Rey Juan Carlos, Spain, Mario Piattini Universidad de Castilla-La Mancha, Spain, Adolfo Sánchez Cronos Ibérica, SA, Spain. Publisher IRM Press Hershey, PA, United States 2002
7. A Strategy for Managing Data Quality in Data Warehouse Systems
1 / IQ01HelfertMaur.pdf
