Business intelligence: a Managerial Approach - Essay Example

BUSINESS INTELLIGENCE SYSTEMS by Student’s Name Code + Name of Course Institution City/State Professor Date Business Intelligence Systems When considering buying a new laptop, a number of activities are involved, the first activity entails defining the decision problem and determining requirements-need recognition (intelligence phase). In this phase, an understanding of the problem is developed (reason for buying a new laptop is established, for instance, the old laptop is experiencing problems). Information about different laptop brands such as Dell, HP, Acer, and Mac is acquired. The second phase entails establishing objectives and goals for buying the laptop, generating different alternatives, determining criteria for purchasing the laptop (i.e. speed, price, and durability) and selecting a method (design phase). During the design phase, the different types of laptops are taken into consideration to determine their quality, prices, and the customer needs. The third phase entails evaluating the alternatives and validating the solutions (Choice Phase). During this phase, information gathered about the different brands plays an important role in evaluating the products in terms of performance, speed, warranty, color, price and durability. The last phase entails implementing the solution (Implementation phase). The implementation phase entails buying the chosen laptop brand, for instance buying a HP laptop because of the high battery life. This phase thus entails making the final choice and effectively purchasing the product. The two tools for visual displays, which may assist in the business performance management, are balanced scorecards and key performance indicators. Balanced scorecard refers to a performance metric which helps to translate the targets and objectives of an organization into a number of actionable initiatives (Turban, Sharda, & Delen et al. 2011). The tool translated the organization mission and vision statement into a comprehensible number of goals and performance measures, which can be quantified and assessed. These measures include various performance categories; financial measures which are supported by the non-financial measures; internal business processes (innovation to build right strategic capabilities); customer (market); and learning and growth (knowledge and skills) and employee performance. Balance scorecards are important in updating or clarifying a business strategy, linking strategic goals to long-term targets and yearly budgets, tracking the main components of the business strategy, facilitating organizational change, incorporating strategic objectives into resource allocation processes (Turban et al. 2011) Key performance indicators are quantifiable measurements, which define and evaluate the success of an organization in achieving its goals. They may be tactical, operation and strategic. They help organizations to define and measure their progress towards organizational objectives. They have to be quantifiable so that they can reflect how well an organization is attaining its stated objectives and goals. They form the basis for business performance management. Meta-data is important in a data warehouse as is helps in managing the mapping between the data warehouse environment and operational environment. Without the mapping controlling the interface is extremely difficult. It also helps to keep track of the changing structure of data over time. It contains the answers to the question about the data stored in the data warehouse- it describes pertinent aspects of the data fully and precisely. It is important to have quality data in a data repository as more reliable and better-informed decisions come from using the right data quality. It is thus important that the data is complete, accurate, and consistent across data sources. The ETL (Extract, Transform, and Load) process can ensure the quality of data in the data warehouse. This process pulls data out of the source systems and places it into a data warehouse. To ensure quality data, ETL involves extracting, transforming, cleansing and loading the data (Turban et al. 2011). The extraction-transformation-loading processes are responsible for extracting the appropriate data from the sources, transporting it to a special purpose area where it is processed and transforming the source data as well as computing new values or records in order to obey the data warehouse structure relation to which they are targeted. In addition, it ensures the isolation and cleansing of the problematical tuples in order to facilitate respect for database constrains and business rules. The final process entails loading of transformed and cleansed data to the proper in the warehouse together with refreshment of its accompanying materialized views and indexes. These processes ensure that quality data is loaded into the data repository. The data warehousing process entails construction and use of data warehouses. The creation of a data warehouse necessitates data cleaning, integration, and consolidation. The use of a data warehouse usually requires a group of decision support technologies (Turban et al. 2011). The data warehousing process brings together information from different sources and puts it into a format, which is beneficial in making business decisions. This objective necessitate various activities such as accurate identification of the business information which is held in the warehouse, identification and prioritization of subject area to be incorporated in the data warehouse, management of the scope of every subject area that will be implemented in the warehouse on an iterative basis. The process also entails development of architecture, which serves as the warehouse technical, and application foundation and categorizing and choosing the software/hardware/middleware elements to implement it. Data is extracted, cleansed, aggregated, transformed and validated to ensure its consistency and accuracy (Turban et al. 2011). In order to support business decision making, the correct level of summarization is defined, also a refresh program, which is in line with business cycles, timing and needs is established. Powerful and user-friendly tools are provided in order to access the data in the warehouse. Text mining refers to the analysis of data in natural language text. Text mining assists businesses to find the “hidden” content in documents, such as added functional associations by mining for context and meaning, relates texts across prior unseen divisions and group documents by general subjects. Text mining techniques are applied to solve various business problems. It helps organizations to derive potentially important business insights from text-based content such as emails, word documents’ and social media postings. Organizations also use text mining techniques in analyzing competitor and customer data to improve competitiveness (Turban et al. 2011). The pharmaceutical industry usually mine research articles and patents in order to improve the discovery of drugs. In academic research, mining and analytics of large sets of data deliver efficiencies as well as new knowledge in diverse areas such as particle physics, biological science and communication and media. Organizations text mine their websites, social media feeds, public blogs, and their documents for business intelligence- they identify emerging trends, explore customer preferences as well as competitor developments. Natural language processing refers to the attempt to mine a full meaning illustration from free text (Turban et al. 2011). It usually uses linguistic concepts, for instance, grammatical structure and parts of speech. It deals with anaphora (“what previous noun does a pronoun or other back-referring phase correspond to”) and ambiguities (grammatical structure and of words, for instance what is being modified by a given prepositional phrase or a given word. The issues associated with natural language processing entail determining if a document has relevance, eliminating commonly used extraneous words (stop-words) such as ‘the’ and ‘a’, reducing words to their Latin or Green roots or stems, resolving homographs and synonyms, allowing for numerous spellings of same word, case sensitivity and punctuation, and calculating frequency or weights of the remaining terms. Neural network are important in a data mining exercise as they have the remarkable capacity to derive meaning from imprecise or complicated data and can be applied to extract patterns and discover trends, which are too complex to be noticed, by either other computer techniques or humans (Turban et al. 2011). Trained neural networks are also experts in the analyzing given information and can be used to offer projections in various new situations of interests and offer response to the “what if” questions. Neural networks are thus important in identifying trends or patterns in data; they are well suited for forecasting or prediction needs and hence their importance in a data mining exercise. The major features of intelligent agents are learning/reasoning, autonomy, reactivity, goal-oriented, communication, cooperation, mobility, and character. They have the capacity to learn from experiences and to adapt its own behavior effectively to the context (Turban et al. 2011). The agent also has to be capable to react well to information or influences from its context. It must have control over its internal states and actions. They have well-defined goals, progressively influence the environment, and so attain its own objectives. They interact within their environment in order to fulfill their tasks and cooperate with other agents and this permits better and faster solutions for complex tasks, which exceed a single agent abilities. They also navigate within the networks and like humans; they are able to demonstrate external behavior with numerous human characters. The intelligent agents’ are applied in information filtering and retrieval and as automated online assistants, scheduling agents, shopping agents, friend-making agents, news watcher, and web document maintenance agents. Speech recognition refers to the ability of a computer system to respond accurately to verbal commands. There are two types of speech recognition-speaker independent and speaker dependent. Speaker independent are devised to recognize all voices and as a result, no training is involved. Speaker dependent softwares work through learning the distinctive characteristics of a single individual voice (Turban et al. 2011). New users have to train the software through speaking to it so that the way a person talks can be analyzed. This technology works in the following way: the users speak into a microphone, the computer applies acoustic analysis to analyze individual sounds (phonemes) uttered. The computer then conducts a search through the available vocabulary database and then selects the words, which seems most likely to be produced. One of the applications of this technology, which may have positive social implications, is its use in school as an assistive device for disabled students. These students may benefit from increase in writing production, improved access to the computer, increased independence, improvements in writing mechanics, improvement in core reading and abilities and decreased anxiety around writing. Speech recognition tool may eradicate potential obstacles for these students, for instance handwriting difficulties. Geospatial technologies such as global positioning system, remote sensing, and geographic information system (GIS) allow mapping as well as analysis of numerous layers of georeferenced data. These technologies are used in analysis, measurement, and visualization of the earth features. Geospatial technologies and data have the capacity to sense, plan for, organize for, counter disasters to protect property and save lives (Turban et al. 2011). They have the capacity to increase government services in areas such as security, health, education, employment and local government services. Geospatial information and data can help the government in supporting decision making regarding these services. It can be used in facilities management, natural resources management, real estate analysis, land use planning, infrastructure and utility planning, environment resource analysis. Through this data, the government is also able to assess the needs of people in various areas and funds that should be allocated to a certain state or region. Essay/Project The project entailed implementing a DSS/BI solution-operational and performance dashboards- for Ballarat Base Hospital to enable would enable quick or real time access to data and help in managing the hospital. The successful design and implementation of a dashboard might be problematic for the hospital and this might cause it to fail to achieve its objectives. The problems that the hospital might experience in the implementation of the dashboards include poorly or unrealistic defined objectives, misalignment of dashboards with business objectives. The hospital might also experience the challenge of defining metric that address desired strategic hospital objectives such as internal quality cost and knowledge management. To ensure that the dashboard will meet expectation, the hospital will have to ensure proper definition of target audiences as well as design of metrics, manageability and simplicity and ensuring continual leadership via a dashboard champion. In addition, the successful implementation of the dashboards and will necessitate a methodology which takes into consideration all the aspects of the projects life cycles. The implementation will also necessitate executive support-the dashboard initiation will require champions-seniour executives who understand the hospital challenges and wield adequate authority to ensure that the dashboards are implemented. They have to stay involved to keep this initiative in the forefront. Without high-level support, the dashboards might fade into obscurity and irrelevance. The operational and performance dashboards will improve efficiency, accelerate decisions, and reduce errors and oversights in clinical practice. They will also play a major role in assisting the management in quality improvement and performance monitoring processes. The dashboards will offer the management the input required to provide effective health services. The operational dashboards will help track every hospital function, which will assist the management, staff, and clinicians to manage, and monitors the flow of patients in the inpatient and emergency department units. The dashboards will also track functions such as human resources, informational technology, and patient relationship management. The dashboards will also support the decision making process of the hospital, as they will easily analyze the overall performance of the hospital, estimate the likely impact upon a hospital ability to deliver medical services, and automatically detect the errors or anomalies. This will help the decision makers to easily identify areas that are doing well and those that need improvement hence making the decision making process fast and easy. Part 2 Performance dashboards support various features that foster decision-making (Turban et al. 2011). They are tailored to support executive meetings that review operations and/or strategy and performance review meetings between managers and the subordinates. The dashboards allow the managers to quickly create or navigate to a desired page and print the output, if desired. They also let the managers and other users to annotate pages or charts, engage in threaded discussions, or kick off workflows to flow through on action items. The features enable the subordinate to explain performance discrepancies and list action steps, and enable executive to review, comment, and approve the action plan. The analysis application in the dashboard also enabled the decision makers to explore data across numerous dimensions and organizational hierarchies. The problems are semi structured as they fall between unstructured and structured problems. Computer support technologies perk up the quality of information on which the decisions are founded, increasing the decision makers’ situational understanding and providing a range of alternative solutions. The system was built by the organization IT department and it is in-house. The users are both the management and other employees. The dashboards use computer graphics to create a visual representation of large collections of information. It utilizes various gauges or displays for different data types and highlights overall metric status using green, yellow or red or similar indicators. Some of the issues encountered entail misalignment of dashboards with business objectives. A data warehouse refers to a larger physical repository of current and historical data (Turban et al. 2011). The data is collected from transactional databases, which contain operational data, it is then organized to offer enterprise-wide, cleansed data in a standardized format and then stored in the warehouse. A data mart refers to a small collection of data, which focuses on a particular function or department or subject. According to Kimball bottom up approach on building a BI solution data marts are created first from operational data to offer reporting and analytic capacities. The focus is on building several dimensional data marts with shared dimension. . Lnmon top-down approach uses relational enterprise-wide approach and the needs of a company as a whole are determined including all the data an organization uses, with the ultimate objective of creating a solution which will cover the needs of all users. ETL (Extract, Transform, and Load) is a four step process which entails extraction of data from the source system, data cleansing and data standardization and the loading of the presentation area data structures from the data staging area. This process helps in detections and removing inconsistencies and errors from the data in order to improve its quality (Turban et al. 2011). High-level architecture major components are the data feeds, the ETL process, the data warehouse, and the presentation layer. A data warehouse is more suitable for business intelligence than a database as it can handle large volumes of data from multiple sources, transform and cleanse the data so that it is constituent and store it in a single place, and present it to the users for analysis and decision-making. The ETL process consists of the extraction, transformation, cleansing and loading stages. Extraction entails reading data from a transactional database, legacy, and OLTP. Transformation entails conversion of the extracted data from its previous form into the form in which it needs to be using commercial or custom-written software called ETL. Cleansing entails ensuring that the staging area if free from errors such as invalid or inconsistent data. Loading entails transferring the data into the data warehouse. This process is important in ensuring that quality data is stored in the warehouse. A data mart can complement the data warehouse as it consists of a smaller collection of data focusing on particular functions, departments, or subjects. This data can supplement the one stored in the data warehouse. Alternatively, organizations seeking particular data on department, subject or function can replace a data warehouse with data marts. Performance measurement helps the management to track the implementation of an organizational strategy through conducting a comparison of actual results against strategic objectives and goals (Turban et al. 2011). Performance management comprises of a closed loop processes, which connect strategy to implementation with the aim of optimizing organizational performance. This is the preferred tool as it sets strategic goals and objectives, established initiative and plans for achieving the goals and monitors actual performance against the set objectives and goals and putting into action the corrective action. The drawbacks of relying solely on financial metrics for measuring performance are financial measures are usually quantitative and tangible and hence they take little account of the intangible and qualitative aspects, they are lagging indicators and they are focused on the short-range and offer limited information on the long term (Turban et al. 2011). Classifications entails identification of categories with predictable and predefined attributed from historical data to predict future behavior using algorithms whereas cluster analysis entails segmenting or grouping to discover new patterns. The intelligent agent help people and act on their behalf, they allow people to delegate work they could have done to the software of the agent (Turban et al. 2011). They can undertake repetitive tasks, summarize complex data in an intelligent manner, and remember things that human beings forget. Artificial intelligence is consistent and thorough, more permanent, less expensive, offers ease of dissemination and duplication, and can be documented. Artificial intelligence can also sense environmental changes and react accordingly. In case a similar change occurs in the future, it is able to refer to the ways it reacted to the past changes to help decide what to do. The disadvantage of artificial intelligence is that it has to be programmed to undertake various tasks unlike natural intelligence Reference List Turban, E, Sharda, R, Delen, D & King, D 2011, Business intelligence: a managerial approach, 2nd Edition, Prentice-Hall, Upper Saddle River, New Jersey. Read More