# Marketing Diploma Course – Question on Unit Eight

Aleksandre Ananiashvili – S12976

Name of student: Name of lecture: Date of assignment: Assignment Title: Unit Number:

Aleksandre Ananiashvili (S12976) Marketing Diploma Course 03/Feb/2010 Answer the Question Unit Eight

Exploratory Data Analysis
Exploratory data analysis (EDA) is an outlook with accompanying techniques that has gained popularity in the last 15 years. John Turkey of Princeton University (Peters 1987) introduced it. It enables to use simple arithmetic and easy-to-draw pictures in order to summarize data (Anderson & Sweeney 2008). This enable to maximize the insights regarding the data, at the same time uncover the potential underlying structures, extract vital variables, detect outliers and anomalies and develop parsimonious models (Verlag 2008). Backhaus stated that the five primary data analysis could be differentiated among the exploratory data analysis techniques, which include factor analysis, cluster analysis, multidimensional scaling, correspondence analysis and neuronal nets (Wolf 2008). These include:  Plotting the raw data (e.g. data traces, probability plots, histograms, lag plots, Youden plots and block plots);  Plotting simple statistics (e.g. mean plots, standard deviation plots and box plots), together with the main effects plots of the raw data;  Positioning plots which can help in order to improve the natural pattern-recognition abilities including the usage of multiple plots per page (Engineering Statistics Handbook n.d.). Most of EDA techniques are graphical in nature with a few quantitative techniques. The primary reason behind this is because the main nature of the EDA. It is to explore in open-minded manner and enables the analyst to have the unparalleled power to do so, which will help in order to entice the data to reveal the structural secrets. At the same time, be ready to gain some new and unsuspected insights about the data (Engineering Statistics Handbook n.d.). Because of this, exploratory data analysis (EDA) procedure is easier to use because there are fewer numerical calculations and formulas to be used and done. The researcher will just have to sort the data values into ascending order and identify the five-number summary. Out of this, the box plot can be constructed (Anderson & Sweeney 2008). Page 1 of 2

Marketing Diploma Course – Question on Unit Eight

Aleksandre Ananiashvili – S12976

In the five-number summary, five numbers are used in order to summarize the data which include: smallest values, first quartile (Q1), media (Q2), third quartile (Q3) and the largest value. This can be done by placing the data in ascending order, in order to know the position or rank of the data and compute the mean. After this, the box plot can be done, this is a graphical summary of data that is based on the five-number summary. A key to the development of this technique is the computation of the median and the quartiles, Q1 and Q3. The interquartile range, IQR = Q3 - Q1 will also be used. This process is applicable in marketing because it will lessen the time and effort to be exerted by the marketing managers and staffs. This can be used in the process of computing and considering the market or the population to be targeted. With the use of this method, the marketing managers will be able to prepare reports and documents that will help to present the result of marketing analysis and study, at the same time, help the viewers or the readers of the report or documents to understand in easier because of its simplicity.

References

Anderson, D R, Sweeney, D & Williams, T A 2008, Statistics for Business and Economics, Cengage Learning. Engineering Statistics Handbook, What is EDA? http://www.itl.nist.gov/div898/handbook/eda/section1/eda11.htm [Accessed 3 February 2010] Peters, W S 1987, Counting for Something: Statistical Principles and Personalities, Springer. Wolf, J 2008, The Nature of Supply Chain Management Research: Insights from a Content Analysis of International Supply Chain Management Literature from 1990 to 2006, Gabler Verlag.J

Page 2 of 2