IS Implementation, OLAP, and Data Mining: Business Analysis and Solutions

Question 1

Question 1

a)Define ‘data’, ‘information’, ‘knowledge’ and ‘wisdom’ and provide a relevant business example for each.

b)Describe the phases of an information system (IS).

c)Define and discuss how an organisation’s business strategy, IT strategy, and IT infrastructure and interconnected.

Question 2

a)State and describe, using the constituency perspective, the four IS solutions traditionally implemented within businesses.

b)Critically discuss how flatness of an organisation’s structure impacts i) transfer of information and ii) formation of information siloes.

Question 3

Describe, with consideration of the business impact, four project management objectives that should be considered when implementing Information Systems.

There are four stages to decision-making, i.e. Intelligence, Design, Choice and Implementation. For each stage, i) explain the output of that stage, and ii) describe an Information Systems solution that might be used to support this stage of decision making.

c)Describe, considering an advantage or disadvantage for each, three alternative software development approaches that could be used to produce software solutions.

Question 4

1. a)  Discuss, by expanding the value chain model, how Information Systems solutions support value creation in businesses.

Expand, in detail, three IS implementation failure factors, providing discussion concerning the implications of each factor, and how each factor might be avoided.

Question 5

1. i)   Define Online Analytic Processing (OLAP), and

(ii) describe, using examples, when OLAP should, and should  not, be used.

(i) Define data mining and describe the information produced, and

(ii) describe, using two examples, how data mining can create business value.

Online Analytic Processing(OLAP) is the technology behind many Business Intelligence applications consists of software tools that are used for data analysis in order to make business decisions and used by the manager, executive, business analyst for reporting purpose. It allows fast multidimensional data analysis that has the capability for manipulating and analysing large volumes of data from multiple perspectives(Babu, 2010). OLAP read optimized to support aggregate queries involving large amounts of data.

Examples of OLAP include: In the finance department, OLAP is used for budgeting or financial performance analysis such as activity-based costing allocation or financial modelling. For instance, financial analyst could use multidimensional data from their previous records to learn what was this year report will look like by using the data from previous years. In sales and marketing departments, OLAP is used for sales analysis and forecasting, marketing research, customer analysis or customer segmentation. For instance, many e-commerce companies when the customers want to buy something will get recommendations for their apparel accessories and other stuff based on their previous purchase history or search.   

OLAP should be used when you want to provide business users with a simple way to generate reports from your company data or when you want to provide a number of aggregations that will allow users to get fast and consistent results.  OLAP apply for aggregate calculations over large amounts of data; where the calculation engine can handle specialized multi-dimensional maths and fast retrieval times e.g. gather the answers from data in less time. It also allows users to segment data into slices that can be viewed into slicing and dicing processes; where it helps users to find trends and discover data without having to know the details of traditional data analysis (Lamount, 2014).

Question 2

On the other hand, OLAP should not be used because it requires organizing data into complicated schemas to implement and administer for examples snowflake or star schema; this means by in OLAP data models tend to be multidimensional, so it is difficult to directly map the entity-relationship where each attribute is mapped to one column. In a single OLAP cube, a large number of dimensions are not allowed and transaction data cannot be accessed with the OLAP system. OLAP is not an up-to-date system because any change in the associate OLAP cube needs an update of the cube and this could be a complicated method(Chin, 2020).  

To implement data mining, first, you need to create a data mining plan by doing business research where you need to understand your project objectives and their requirements. After you collected the relevant information, you proceed to the next step which is Data Quality Checks; this ensures that there are no bottlenecks in the integration process and it helps to spot missing data before undergoes mining. Then the next step is data cleaning or preparation where it preparing the data in the right format e.g. to check duplicate data and anonymizing before mining. After final checked, then now Modelling by using algorithms to identify patterns with data based on several conditions (Sharma, 2020).

It is normally divided into two categories which include Predictive tasks and Descriptive tasks. Predictive tasks are when the business predict the value of a particular attribute based on the value of other attributes. The attributes that used to predicted include target or dependent variable, while explanatory or independent variables are used for making the prediction(Pang-Ning et al., 2019, p29). Predictive modelling goes deeper to classify events in the future or estimate unknown outcomes, for examples, using the Regression technique to measure the strength of the relationship between a series of independent variables to one dependent variable or using decision trees to identify a probable occurrence in each branch.

Descriptive tasks are when the company use to derive patterns that summarize the underlying relationship in data, they are often exploratory in nature and frequently require postprocessing techniques to validate and explain the result(Pang-Ning et al., 2019, p29).  Descriptive modelling groups previous or historical data to determine the reason behind success or failure by product preferences.  Clustering can be one of the techniques, cluster analysis seeks to find groups of closely related observations in order to group similar records together.  Forecasting can also be used to forecast what other value will be by using an existing series of values to predict.

Data mining may be used to discover patterns or trend of your current business and it can create business value by helping them to create more informed decision and increase the demand for Customer relationship management(CRM). Data mining could help to increase company revenue and increases business optimization. It also can help to detect fraud, if the business is a technology company like an apple; data mining can help introduce a new product (Tawde, n.d.)

EXAMPLES

Anomaly detection: identifying observations whose characteristics are significantly different from the rest of the data. Anomaly detection mainly discovers the real anomalies and avoid falsely labelling normal objects as anomalous e.g. detect fraud or unusual patterns of your business(Pang-Ning et al., 2019, p33).   

Example: Credit card company, the company uses personal information of cardholder to check legitimate before making each transaction. Since the number of fraudulent cases is relatively small compared to the number of legitimate transactions, anomaly detection techniques can be applied to build a profile of legitimate transactions for the users. When each new transaction arrives, this technique is checked and compared against the profile of the user. If the attributes of the transaction are very different from the previously created profile, then the transaction is flagged as potentially fraudulent.

Association analysis: is used to discover patterns that describe strongly as- sociated features in the data. The discovered patterns are typically represented in the form of implication rules or feature subsets(Pang-Ning et al., 2019, p32). It extracts the most interesting patterns in an efficient manner and searching groups that have related functionality.