Methods of Clustering Data in SPSS- Nandan Shah (12148)

Clustering is a method of identifying groups of individuals or objects that are similar to each other but different from individuals in other groups can be intellectually satisfying, profitable, or sometimes both.

Clustering helps to take advantage of these similarities to target offers to subgroups that are most likely to be receptive to them thereby making the process more effective & cost efficient.

SPSS has three different procedures that can be used to cluster data:

• Hierarchical Cluster

• K-Means Cluster

• Two-Step Cluster

Hierarchical Cluster

Hierarchical clustering is used if the data set is small & one wants to easily examine solutions with increasing numbers of clusters.

Hierarchical clustering is one of the most straightforward methods. It can be either agglomerative or divisive.

Agglomerative hierarchical clustering begins with every case being a cluster unto itself. At successive steps, similar clusters are merged. The algorithm ends with everybody in one single cluster. It is a 'bottom up' approach. Divisive clustering starts with everybody in one cluster and ends up with everyone in individual clusters. It is a 'top down" approach.

In agglomerative clustering, once a cluster is formed, it cannot be split; it can only be combined with other clusters. Agglomerative hierarchical clustering doesn’t let cases separate from clusters that they’ve joined.

To form clusters using a hierarchical cluster analysis, one must select:

· A criterion for determining similarity or distance between cases

· A criterion for determining which clusters are merged at successive steps

· The number of clusters you need to represent your data

Similarity or Distance Between Cases

Because the goal of this cluster analysis is to form similar groups one has to decide on the criterion to be used for measuring similarity or distance. Distance is a measure of how far apart two objects are, while similarity measures how similar two objects are. For cases that are alike, distance measures are small and similarity measures are large. There are many different definitions of distance and similarity. Some, like the Euclidean distance, are suitable for only continuous variables, while others are suitable for only categorical variables. There are also many specialized measures for binary variables.

Standardizing the Variable

If variables are measured on different scales, variables with large values contribute more to the distance measure than variables with small values. Hence the variables need to be standardised.

How Should Clusters Be Combined?

Agglomerative hierarchical clustering starts with each being a cluster. At the next step, the two items who have the smallest value for the distance measure (or largest value if you are using similarities) are joined into a single cluster. At the second step, either a third case is added to the cluster that already contains two cases or two other cases are merged into a new cluster. At every step, either individual cases are added to existing clusters, two individuals are combined, or two existing clusters are combined.

K-Means Cluster

K-Means clustering doesn't require computation of all possible distances unlike in the case of Hierarchical clustering that requires a distance or similarity matrix between all pairs of cases. Thus if the number of clusters is known & the data is moderately sized, one can use k-means clustering.

K-Means clustering differs from hierarchical clustering in several ways. One has to know in advance the number of clusters one wants. One can’t get solutions for a range of cluster numbers unless one reruns the analysis for each different number of clusters. The algorithm repeatedly reassigns cases to clusters, so the same case can move from cluster to cluster during the analysis. In agglomerative hierarchical clustering, on the other hand, cases are added only to existing clusters. They’re forever captive in their cluster, with a widening circle of neighbours.

The algorithm is called k-means, where k is the number of clusters required, since a case is assigned to the cluster for which its distance to the cluster mean is the smallest. The action in the algorithm centers around finding the k-means. One starts out with an initial set of means and classify cases based on their distances to the centers. Next, step is to compute the cluster means again, using the cases that are assigned to the cluster; then, reclassify all cases based on the new set of means. This step is repeated until cluster means don’t change much between successive steps. Finally,the means of the clusters are calculated once again and the cases are assigned to their permanent clusters.

Initial Cluster Centers

The first step in k-means clustering is finding the k centers. This is done iteratively. Start with an initial set of centers and then modify them until the change between two iterations is small enough. If one has good guesses for the centers, one can use those as initial starting points; otherwise, one can let SPSS find k cases that are well-separated and use these values as initial cluster centers.

After the initial cluster centers have been selected, each case is assigned to the closest cluster, based on its distance from the cluster centers. After all of the cases have been assigned to clusters, the cluster centers are recomputed, based on all of the cases in the cluster. Case assignment is done again, using these updated cluster centers. One has to keep assigning cases and recomputing the cluster centers until no cluster center changes appreciably or the maximum number of iterations (10 by default) is reached.

Final Cluster Centers

After iteration stops, all cases are assigned to clusters, based on the last set of cluster centers. After all of the cases are clustered, the cluster centers are computed one last time. Using the final cluster centers, you can describe the clusters.

Two-Step Cluster

If the data file is large or a mixture of continuous and categorical variables, SPSS two-step procedure can be used. It requires only one pass of data (which is important for very large data files), and it can produce solutions based on mixtures of continuous and categorical variables and for varying numbers of clusters.

The Two-Step Cluster Analysis procedure is an exploratory tool designed to reveal natural groupings (or clusters) within a data set that would otherwise not be apparent.

The algorithm employed by this procedure has several features that differentiate it from traditional clustering techniques:

– The ability to create clusters based on both categorical and continuous variables.

– Automatic selection of the number of clusters.

– The ability to analyze large data files efficiently.

Step 1: Preclustering: Making Little Clusters

The goal of preclustering is to reduce the size of the matrix that contains distances between all possible pairs of cases. Preclusters are just clusters of the original cases that are used in place of the raw data in the hierarchical clustering. As a case is read, the algorithm decides, based on a distance measure, if the current case should be merged with a previously formed precluster or start a new precluster. When preclustering is complete, all cases in the same precluster are treated as a single entity. The size of the distance matrix is no longer dependent on the number of cases but on the number of preclusters.

Step 2: Hierarchical Clustering of Pre-clusters

In the second step, SPSS uses the standard hierarchical clustering algorithm on the preclusters. It takes sub-clusters resulting from the first step as input and then groups them into the desired number of clusters. Since the number of sub-clusters is much less than the number of original records, you can use traditional clustering methods effectively. SPSS uses the agglomerative hierarchical clustering method primarily because it works well with the auto-cluster procedure.

References:

http://www.resample.com/xlminer/help/HClst/HClst_intro.htm

http://en.wikipedia.org/wiki/Cluster_analysis

http://www.spss.ch/upload/1122644952_The%20SPSS%20TwoStep%20Cluster%20Component.pdf

Submitted By:

Nandan Shah
Roll Number: 12148
Finance Batch of 2009-11
SIBM Bangalore.

Use of Discriminant analysis in Credit Risk Models

Discriminant analysis is a technique used for classifying a set of observation into groups or classes. It is widely used in finance particularly in loan or credit appraisal.

Credit scoring model assigns a numerical value to a firm which indicates whether a firm is likely to default or not. It is used to assess the creditworthiness of a firm for giving approval of loan in a short span of time.

Some of the models that use Discriminant analysis are

1. Fisher Linear discriminant Analysis
2. Parametric discrimination
3. K-nearest neighbor

Fisher Linear Discriminant Analysis segregates a larger group into homogeneous subgroups. The larger group could be potential borrowers and subgroups could be good and bad borrowers. Altman’s Z score is a famous example of discriminant analysis.

In Altman Z Score the equation use to calculate score is as follows

Z = 0.012T1 + 0.014T2 + 0.033T3 + 0.006T4 + 0.999T5.
Where
T1 measures liquidity, T2 measure profitability, T3 measures operating efficiency, T4 measure market fluctuation of equity and liability and T5 measures sales turnover.

Based on the score (Z score) calculated from the above equation, new borrowers are classified into default or non default subgroups. This model is highly successful in manufacturing sector and is extensively used by banks and other institutions before the sanctioning of loans. Variant of this model is also developed for other sectors. The greatest advantage of this model is its simplicity; it uses only few balance sheet items and requires a minimal time and resources to compute. The basic concept behind this model is Linear Discriminant Analysis.

In Parametric discriminant approach, a score function is used to determine the members of the subgroup. It uses regression, Logit or other statistical techniques for analysis. Whether the value of the score falls above or below a certain threshold determines which subgroup the observation is placed.

In K-nearest neighbor approach, properties of firms that have already fallen into the categories of interest and categorizes a new entrant by how closely it resembles the members already in each of the groups. It is simplest of all machine learning algorithm in which an object is classified by a majority vote of its neighbors, with the object being assigned to the class most common amongst its K nearest neighbors. This method uses regression by simply assigning the property value for the object to be the average of the values of its k nearest neighbors. It can be useful to weight the contributions of the neighbors, so that the nearer neighbor contributes more to the average than the more distant ones.

The approach is depicted in the above figure. The green circle is taken as the sample and the elements in the inner circle are its nearest neighbor. Circles depict the clusters or groups segregated on the basis of the resemblance its test sample.

Various propriety and freely available software are available that provides facilities for discriminant analysis. Some of the softwares that provide discriminant analysis facilities (partial or complete) are SPSS, business objects, cognos and MS Excel.

Sources

1. Measuring Credit Risk and Managing Credit Risk by Arnaud de Servigny and Olivier Renault.
2. http://en.wikipedia.org/wiki/Altman_Z-score
3. http://en.wikipedia.org/wiki/Discriminant_analysis
4. http://en.wikipedia.org/wiki/K-nearest_neighbor_algorithm

Submitted By:
Mrityunjay Kapoor
Roll Number: 12147
Finance Batch of 2009-11
SIBM Bangalore.

Using Discriminant Analysis to Predict Financial Distress – Mansi Dubey (12092)

This article is a perfect example of how discriminant analysis can be used in the field of economics and finance. A study conducted by Benzing Cynthia, published on August 1, 2000 uses discriminant analysis and regression to determine which macroeconomic variables are predictive of financial crisis in emerging countries.
Emerging countries were divided into two groups:
 The first group of crisis countries had entered into either a stand-by agreement or an extended agreement with the International Monetary Fund between 1995 and 1999.
 The second group of emerging countries, matched by region with the first group, consisted of countries who did not experience financial problems.
The discriminant analysis tested 18 variables related to liquidity, debt levels, current account deficits, and monetary policy to determine which variables were significantly different between the two groups. Then, with a dummy-dependent variable for crisis and non-crisis countries, regression was used to determine which variables had the highest joint predictive value.
The study found that the bank reserves to demand deposits ratio, exports to imports ratio, the percentage of growth in domestic credit, and the government spending to gross domestic product (GDP) ratio had predictive value.
Countries that faced financial distress had a higher bank reserves to demand deposits ratio and percentage of growth in domestic credit 12 months before financial distress. These same countries had a lower exports to imports ratio and government spending to GDP ratio.

Ref: http://www.allbusiness.com/finance/635773-1.html

Name: Mansi Dubey
Roll Number 12092
Finance Batch of 2009-11,
SIBM Bangalore.

SPSS panacea to lot of organizational problems - By K Sailesh (12026)

Why SPSS…
Ever wondered why successful firms are using SPSS for their business intelligence decisions. Read on and if you are skeptical about using SPSS for your decision making let me give you some I guess few reasons which I think would be enough to remove all the misconceived notions about this wonderful piece of analysis tool and make your nerve cells crave for SPSS.

WHAT MAKES SPSS UNIQUE!!!
In no particular order because I guess every reason stands out!!
WORLDWIDE LEADER IN ANALYTIC (PREDICTIVE) TECHNOLOGIES:
Ask big companies and I’m sure they would vouch for SPSS being their most preferred business analysis tool in their research department. For over 37 years now it has been with over a client base of in excess of thousands, be it academic institutions, governments, health care providers, market research companies it has enabled organizations to focus and improve their performance.
Who are the people who use SPSS : Quoting numbers as that what justifies I guess, since we all need number crunching in this course: around 50 US state governments base their decisions on it, the top 10 MR firms of the world, too many universities in the US (aroud 90 percent ), top 10 pharma companies and off course you have the largest news paper companies too

SOLVING ORGANIZATION CHALLLENGING ISSUES:
Academic institutions use SPSS to manage resources, predict demand for their courses and be able to able to raise funds effectively
Governments as explained above use SPSS. Now let me tell you how – lot of frauds happen, so many scams you hear of , they can use this software to detect frauds, whether people are complying or not with the rules and regulations, and offcourse for public safety and homeland security.
Even non profit org want a share of the cake. They are able to raise funds by taking accurate decisions based on the analysis they do on you know what – SPSS ofcourse.
So what are you waiting for ??
If not yet satisfied let me give you another!!
Everything happens real time In organization – so the other advantage of SPSS is:
REAL TIME DECISION OPTIMIZATION:
Firms interact with customers over the web, through the phone (thanks to Mr BELL), through direct mail or through the store, you got to take decisions that will satisfy the customers. Take a wrong one and he will spread real bad about you unfortunately to so many people (you got to hope that he doesn’t know too many people – one of the nonsense of viral marketing) . Why do you got to be in such a position. With SPSS, the wonderful thing gives you recommendations, along with supporting information on the most appropriate action to be taken with a particular customer. The software can help you predict alternatives, what a customer is likely to respond to. And of course what do you get out of it – you are able to achieve your organization goals.
And now a list as to why you need SPSS :
- SPSS can work with your existing applications – So don’t bother investing a lot on other software that not only cost lots but cost even more to maintain them. SPSS can be integrated with your existing applications
- It doesn’t take long to implement SPSS and make it running for your business decisions
A bit of statistic being thrown here:
BankFinancial, an 80+ year-old financial institution with more than $1.5 billion in assets, implemented an SPSS predictive analytics solution designed to optimize marketing campaigns. Within two months, the solution was up-and-running and the organization experienced a 700% increase in campaign response rates.
- You can easily learn it and implement it along with CRM
Another example:
FBTO, an insurance company and part of the Achmea Group, one of the largest financial services providers in the Netherlands, implemented a predictive analytics solution from SPSS. In two months, FBTO achieved measurable increases in marketing response rates and lower overall marketing costs.
- It robust and expandable
So you see be it learning SPSS in the classroom or it being used in companies I’ m sure by now you must have guessed what you are missing if you aren’t smart enough to use SPSS
By the way like everyone says SPSS started with social science and now can be used for making millions of dollars (billions too.. ) if you use it.
I leave you with this:
“it is possible to lie with statistics, but is impossible to tell the truth without it” – INDIAN STATISTICAL INSTITUTE

Source: http://www.spss.com/corpinfo/spss_edge.htm#1

Chi Square Analysis: Few insights and glances of some of its applications in the industry in Marketing; Some key learnings from the statistic : Gargi Ghosh (Roll: 12135)

Insights on Chi Square Statistic:
 The chi-square test is a non parametric statistical test used to examine differences with categorical variables. Chi-square statistics use nominal (categorical) or ordinal level data, thus instead of using means and variances, this test uses frequencies.
In SPSS, it can be easily found out by the sequence: Analyse --> Descriptive statistics --> Crosstabs. 

The value of the chi-square statistic is given by: c² _{= S (f observed - f expected)}² _{/ f expected,} where
_{f observed} is the observed frequency and _{f expected} is the expected frequency.
The expected frequency is calculated by: _{f expected = (row total * column total) / total frequency}

_{and, the degree of freedom is determined as: c}² _{df = (no. of rows – 1)*(no. of columns – 1)}

Generally the chi-squared statistic summarizes the discrepancies between the expected number of times each outcome occurs (assuming that the model is true) and the observed number of times each outcome occurs, by summing the squares of the discrepancies, normalized by the expected numbers, over all the categories.

Data used in a chi-square analysis has to satisfy the following conditions:

1. Randomly drawn from the population,
2. Reported in raw counts of frequency,
3. Measured variables must be independent,
4. Observed frequencies cannot be too small, and
5. Values of independent and dependent variables must be mutually exclusive.

There are two types of chi-square test for two similar but distinct circumstances:

·         The Chi-square test for goodness of fit which compares the expected and observed values to determine how well an experimenter's predictions fit the data.

·         The Chi-square test for independence which compares two sets of categories to determine whether the two groups are distributed differently among the categories.

Goodness of fit means how well a statistical model fits a set of observations. A measure of goodness of fit typically summarizes the discrepancy between observed values and the values expected under the model in question. Such measures can be used in statistical hypothesis testing, e.g., to test for normality of residuals, to test whether two samples are drawn from identical distributions.

The chi-square test for independence is used to determine the relationship between two variables of a sample. In this context, independence means that the two factors are not related.

Applications in industry:
Let us consider the following decision situations for a Marketer:

 1) Are all package designs equally preferred?

2) Are all brands equally preferred?

3) Is there any association between income level and brand preference?

4) Is there any association between family size and size of washing machine bought?

5) Are the attributes educational background and type of job chosen independent? 

The answer to these questions requires the help of Chi-Square (c²) analysis.

The first two questions can be solved using Chi-Square test of goodness of fit for a single variable.

The solutions to questions 3, 4 and 5 need Chi-Square test of independence in contingency table.

Some Learnings:
 
Some disadvantages of chi-squared tests are that its value depends upon the degree of freedom of the crosstabs and the sample size. And it cannot detect the direction of an association. So one has to be careful while interpreting its value when comparing two different crosstabs with different degrees of freedom or with very different sample sizes.

 Chi-square distribution is a skewed distribution particularly with smaller degree of freedom (d.f). As the sample size, and therefore the d.f. increases and becomes large, the Chi Square  distribution approaches normality.

Chi Square Analysis is widely applied in a number of service industries, like, Retail, banking, Airlines, Healthcare, Education where a huge amount of data are dealt with and in most of the cases, the data are highly inter related with each other.

P.S.: Few of the above insights are adopted from: http://cnx.org/content/m13487/latest/

The Big Picture - Gomathi Shankar K 12022

SPSS and Business Intelligence are two new words added to my MBA dictionary today. Words, being as deceptive as they are, usually hide a big picture behind them. For instance, 'An alley down the 4th street of Jaya Nagar, Bangalore' is a bigger word description than 'The Sahara'. But to understand the magnitude of the words, one has to see the big picture, philosophically and in this case of Jaya Nagar Alley vs Sahara, quite literally too!


So let me try to depict the big picture behind these two words. SPSS forms a subset of Business Intelligence Tools which in turn forms a subset of Data Mining Applications which in term forms a subset of Statistics Applications which is quite big enough a picture for us. So let us stop here and explore this picture in front of us. For starters, some Wiki definitions as follows:


Data Mining: Data mining, a branch of computer science and artificial intelligence, is the process of extracting patterns from data.


Business Intelligence: Business intelligence (BI) refers to computer-based techniques used in spotting, digging-out, and analyzing business data, such as sales revenue by products and/or departments, or by associated costs and incomes.


SPSS: A computer program for statistical analysis, original designed for Social Sciences. 


An interesting example of an application of Data Mining that I spotted while mining some websites is this article, which tells about a unique application that detects spoilers in movie reviews and allows for readers to confidently read online reviews without the fear of general knowing who committed the murder. This application considers combinations of words as data and predicts spoilers using certain combinations of words and some linguistic logic. That is data well mined and well applied! Not in business strictly but this example may tell us that data is everywhere is gold is everywhere under it to be mined. 


In Business Intelligence, data usually takes various forms for various purposes. In general, all these efforts are to answer one basic question "How well do you know your customer"? Experienced businessmen will admit that it took years to find answers to this question, and that too with a touch of uncertainty well masked. Such is the complexity of this single question. It does pose many branch questions which give clues to aid analysts to grope in the dark till they clutch the root. 

Every row and column there can hide fascinating mysteries behind. Know which door to open.  SPSS will open it for you.

SPSS is one intelligent pair of goggles used by the analyst to aid his search for truth/insight in the dark. As they are always, truth and insight only offer themselves to the deserving. So not all who wield the wand can do magic with SPSS or any other tool, because this tool is an intelligent servant who will do what is asked to be done and nothing more than that. Only the trained and experienced will know what to ask from this genie. There lies the route map to the root question mentioned above. As is said often, in business, one who has the right answers will always be a good employee and one who has the right questions will always be his master. SPSS can only be mastered with the understanding and the knack of finding the right question. 


From the brief discussions we had in class today in the retail data, we ordered SPSS to show us the what, who and where of customer satisfaction in several lights. A sample below:


What: Customer dissatisfaction in service
Where: Store 2
Who: Female customers
Where exaclty: Clothing section in Store 2


'Why' is the ultimate question to be answered. After getting to know the above answers, the analysts throw several hypotheses to find the answer for the elusive 'Why'. 


We also discussed about spurious associations in class today. Some people may remember Sania Mirza's skirts more vividly. What it emphasizes is that data, as obedient a server to a wise master, can also be as good a misleading conundrum to the hazy eyed questioner. Hence, it is always essential to check, re-check and cross-check to believe in a hypothesis. 


Cross tabulation and frequencies, the specifics of today's discussions become still smaller in the larger picture as a subset of SPSS too. They are only descriptive methods, or in simple words, nothing but adjectives for data or a customizable display panel of data. What we did today was to learn to adjust our eyes to this display panel and ask what we want to see. 


They will only direct us to hypothesis generation. Conclusive evidences on hypotheses can be arrived after some more playing with data. They are like night vision goggles. They are tools designed to show what can not be normally seen in the clutter. We have learned to adjust our eyes to these goggles today. For example, we were able to focus our sight to only to a negative customer contact in the clothes section of Store 2 today with these tools from a mountain of data. Who knows, there might be an irritating employee in that store section who is getting on the nerves of customers there. The answer is yet to be found out with conclusive evidence. Let us try to convict this black sheep conclusively by subjecting him to cross examination with other tools in the coming days. 


Focus on the big picture!


- Gomathi Shankar K

A summary of day 1 of the SPSS workshop : Padmini Krishna 12152

The first class of the SPSS workshop exposed us to the very basics of SPSS; its usage, origins and some of its applications.

Firstly, we were familiarized with the SPSS software in terms of its usage and learnt the terminology associated with it like the variables, labels, missing values etc.

We were then taken through an example which encapsulated the statistical concepts of frequency and cross tabulation.

Frequency:  is the number of occurrences of a particular event in an experiment.  Frequency analysis is thus a univariate method of analysis.

Cross tabulation: is a method of analysis which establishes an interdependent relationship between two tables of data. Since it involves two data sets, it is a bivariate method of analysis. The general procedure for running a cross tabulation is:

- Establish a null hypothesis- we begin by saying that “there is no relationship between X and Y”

- Run the cross tabulation analysis- analyse the relationship between X and Y

- Run the Chi-square analysis- establish if there is in fact a relationship between X and Y.

This method of analysis finds many applications in Market Research. For example, if a supermarket wants to understand the purchase pattern of a certain segment of its customers, it can cross tabulate the income levels of its customers against the frequency of purchase. 

Another example is that a clothing brand can cross tabulate the gender of the customers against the colours they pick. This will give the company an insight into how the colour preferences vary according to the gender.

We might be tempted to immediately draw conclusions based on the results of the cross tabulation, however, we were told to desist till we get the result of the Chi-Square analysis. The Chi-square analysis is a test of independence. It is that which ultimately proves the existence (or not) of a relationship between two variables. It is that which separates fact from conjecture.  The result which is in terms of a number up to two or three decimal places , signifies the level of confidence in the hypothesis. Thus a result of 0.05 signifies a 95% level of confidence. 0.05 is a popularly accepted benchmark although it can vary depending on the criticality of the decision. A result of less than 0.05 says that our null hypothesis is bunk or in more polite words, “wrong”. A result of over 0.05 allows us to pat ourselves on the back for our accurate hypothesis.  

Thus, we learnt that SPSS, a tool that was originally intended to be of use in the arena of social sciences is now an indispensible tool for market researchers, health researches and government agencies to name a few. When combined with pragmatic marketing sense, it can help glean meaningful insights into consumer behaviour which can translate into monetary benefits and growth of the company. 

BUSINESS ANALYTICS... ...The power of knowing what’s next! Sweta Jobanputra 12170 HR

In today’s ever changing market, especially the Indian market, leveraging the power of technology and remaining one step ahead of the competition is one of the key motto to success. Applying intelligent information systems and how it could help in making transformational decisions is the talk of success these days. Here comes the power of Business Analytics and how it helps in decision making through predictive analytics.

Predictive analytics is a relatively new concept to many in the software market. So what is Predictive Analytics? “Predictive Analytics is about taking effective action based on data that is used to draw reliable conclusions about current as well as future events”. So now the question is how does a company know if it’s abilities fit into the Business Analytics background. To start off with, like any other company using Business Analytics capabilities, the primary question is to report “what happened”. Using various software tools such as SPSS or even SAS, the company performs analysis to dig out the route causes of “why it happened”. Moving on, the managers then monitor on what is happening now and whether or not any change is required, and then the final step helps the company “predict” about “what might happen”. If a company successfully maps these questions to Business Analytics capabilities, using the software mentioned above, then its easy to analyze what functionalities can help the company address and answer the four key questions.

Since I have extensively mentioned technology, we’ll now see what the technologies that make up predictive analytics are. Applying software such as SPSS helps in predicting just about anything, be it the current market scenario or how the customer trends are going to change in the future. All one needs to do is to identify the variables that would be key considerations for making the decisions, collect data specific to each variable through surveys, questionnaires etc, enter the data into the software and then analyze using various statistical techniques.

Analysing through SPSS and reaching a conclusion about a decision, is just about making a few permutations and combinations of one or more than one variable. For example, using the command “Frequencies” one can get a count on how many people, or more specifically how many women are satisfied with the clothing of a particular store. And what more? The satisfaction levels of the women buying clothes from store 1 v/s store 2 can be analysed as well. This could help the company actually identify whether or not the audience is satisfied with the store or not, and if not, what is the problem and how can it be solved.

Imagine using the tradition approach of collecting data and reporting manually, and then finally making the decision, by the time the decision has been arrived, the market has moved forwards. This is where using technology and softwares such SPSS helps in quick decision making, and we all know, that is what we need to cut through the competiton.

Hierarchical Clustering-Problems & Limitations - By, Bipin Easow HR-12094

Hierarchical clustering is used when a underlying application needs a hierarchy. Such kind of a cluster produces high quality clusters. However there are many problems and limitations with the clustering technique which are as follows :

Ability to handle different cluster sizes

Hierarchical clustering can face problems in handling the relative sizes of the clusters that are merged. There are two approaches : weighted, which treats all clusters equally and unweighted, which takes number of points in each cluster into account. Here the number of weighted and unweighted refers to the data points not the clusters. In other words treating clusters of unequal size equally gives different weights to the points in different clusters, while taking cluster size into account gives points in different clusters the same weight. This is one major limitation of the hierarchical clustering algorithm.

Merging Decisions are Final

Agglomerative hierarchial clustering algorithms tend to make good local decisions about combining two clusters since they can use information about the pairwise similarity of all points. However, once a decision is made to merge two clusters, it cannot be undone at a later time. This approach prevents a local optimization criterion from becoming a global optimization criterion.

Cost Factor

Hierarchial clustering algorithm is expensive in terms of its computational and storage requirements. The fact that all merges are final can also cause trouble for noisy, high dimensional data such as document data.

In turn these problems can be addressed to some degree by first partially clustering the data using another technique called K-means.

Introduction to SPSS – Day 1 of the SPSS Business Analytics Workshop; Abhishek Nair (12063)

SPSS is a statistical analysis program, commonly used in Research, whether it be in the field of Social Sciences, Marketing, Operations, or any other field where it is desirable to extract meaningful connections from large sets of data.

SPSS was initially developed for use in the Social Sciences, as evident in the name – Statistical Package for the Social Sciences.  (Source: http://en.wikipedia.org/wiki/SPSS)

Day 1 of the classes focused on a basic introduction to SPSS and some of its data-crunching capabilities.

Interface:  The interface consists of a spreadsheet-like screen (GUI similar to Excel), which is used for entry of data, and an Output Screen (CLI/GUI), in a separate window.

There are 2 tabs at the bottom, named Data View and Variable View – to enter data. All outputs in terms of calculations and analyses performed on this data are shown in the Output Window.

Data Entry: Data type, along with Name of variable, length, labels, measurement type, etc. is defined in the Variable View, whereas actual entries of respondents, as entered by the Analyst, are visible in the Data View. An instance of a variable, i.e., an actual entry, is called a Case.

Data Analysis:  Once basic data definition and entry are complete, the Analyst is ready to use the capabilities available at his disposal, thanks to SPSS, to crunch this data.

Some tools available in SPSS are a set of descriptive statistical tools, called Cross Tabs.

Cross tabs allow a researcher to check the dependence between any two variables.

Examples: An example for this was illustrated, using a General Social Survey (With 67 variable types), built into SPSS. The sample was from a Social Survey conducted in the US, and contained respondents answers to such questions as Age, Gender, Race, Education Levels, Personal Life, Income, Political Views, Music tastes etc.

The Hypothesis tested was that people who get married below a particular age (as indicated by Age of First Marriage) have lower education levels, and the corollary of the same is that people with higher education levels get married later in life.

The data was categorized based on education level of individual respondents, as well as when they first got married.  A chi-square analysis was conducted on this data, and since the confidence level (indicated by a score of .005 or more) was above 95%, it was accepted.

Hence, people with lower education levels had their first marriage earlier.

Similarly, a Retail Satisfaction Survey (With 18 Variable Types, ranging from Age, Gender, Payment methods, to satisfaction levels) was also analyzed, wherein data indicated that the general satisfaction level of shoppers who visited the Clothing Dept. in a Store (Store 2 of 4)of a Retail chain, who were attended to by staff , was actually lower than those who were not attended to.

It followed that most of these shoppers were also dissatisfied in the Shoe and Stationery Depts. Indicating they were women. This was corroborated by incidental data collected by the store (Age, Gender etc., apart from Satisfaction Level Indicators).

The second data set, i.e., that of the retail satisfaction survey, exposed us to the concept of spurious data. The data may look alright, even when subjected to rigorous statistical analysis. However, it is up to the researcher to use his experience as well, to ensure that the inferences are not erroneous in nature.

Conclusion: SPSS is very powerful statistical software, with a large number of mathematical and statistical functions, which places significant data analysis capabilities in the hands of the present-day researcher. A simple peek into the software on Day 1 was sufficient to convince me about the relevance and power of this software.

As a student of Marketing, as well as a future Marketer, I am left with no doubt that this tool will provide me with significant insights into the minds of the consumer. Any manager must have rudimentary knowledge of this program.

A word of caution: However, this tool is only as good as the User, and it is ultimately up to the experience and skill of the Manager to decide whether or not to risk his scarce resources on the results of a Computer Program alone.

Remember, torture numbers and they will confess to anything.

-

Abhishek Nair,

Roll Number 12063,

Marketing Batch of 2009-11,

SIBM Bangalore.