Alberto Rossin Expert statistician and skilled teacher

I can help you to segment your target market, determining the differences in purchasing behavior of your customers, from several nationalities, cultures, ages, social strata, etc. The knowledge of these features can allow marketers to design more accurately advertising campaigns or to make more profitable promotional offers, tailored to the different segments of buyers. This is an example of application of factor analysis to a sample of English, Chinese and Italian customers to whom a questionnaire, which aimed to establish their preference in the purchase of luxury goods, had been administered. The data processing was done with the IBM SPSS software. The Factor Analysis reduced all the preferences to a small number of factors that determined most of preference’s variance. The analysis describes some luxury goods customers’ profiles, based on their shopping preferences that allow a first market segmentation In the following tablee the factors I found are resumed (using Principal Components extraction method with Equamax rotation method), with the related percentages of variance explained. ITALIAN BRITISH CHINESE Factor 1 35% 33% 32% Factor 2 15% 19% 21% Factor 3 9% 10% 11% Factor 4 7% 8% Factor 5 7% The first Factor Analysis rotated matrix allowed me to describe the Italian luxury goods customers as: 1. Factor: customers who prefer shopping in store because they want to feel and touch products that are immediately 2. Factor: customers who prefer shopping in store. They enjoy being attended to by a sales person, the store’s atmosphere and the possibility to access limited editions that cannot be found online. These customers don’t think that the store experience can be replaced by the online. 3. Factor: customers who buy luxury goods products in store and online are likely to continue purchasing through both channels 4. Factor: customers who prefer shopping online to access a wider selection of products and to read peers’ reviews. They are likely to buy luxury goods products online in the future. The second Factor Analysis rotated matrix led me to describe the British luxury goods customers as: 1. Factor: customers who bought online and in store in the last six months and are likely to continue buying through both channels. 2. Factor: customers who prefer shopping online as they have limited time and want to access a wider selection of products. 3. Factor: customers who buy in store to feel and touch the products that are immediately available. 4. Factor: customers who don’t think the store experience can be replaced by the online. They enjoy the store atmosphere and being attended. 5. Factor: women who like browsing online but then prefer purchasing in store as they are attracted by the possibility to access limited editions. The third Factor Analysis rotated matrix allowed me to describe the Chinese luxury goods customers as: 1. Factor: customers who bought luxury goods online in the past six months and, even though they visit the store they don’t like to be attended to by a sales person and prefer purchasing online. They are likely to buy online in the future. 2. Factor: customers who buy in store because they like to experience the products, they enjoy the store atmosphere and appreciate the possibility to access exclusive editions not available online. 3. Factor: customers who have predominantly bought in store in the past six months and are very likely to continue shopping in store in the future. They don’t think that the store experience can be replaced by the online one.

This analysis, carried out for an Italian customer, concerns the psychometric properties of the Reason for Living Inventory (RFL) of Linehan, Goodstein, Nielsen and Chiles (1983) in its Italian version. This study was performed on a sample of 200 subjects, interviewed with the questionnaire RFL My contribution to the research consisted in performing appropriate statistical analysis for the research objectives, which were to assess the following aspects: - Internal consistency of RFL scale applied to the Italian, through measures of reliability (Cronbach's alpha) - Differences in the "reasons to live" between the experimental and the control group (with the Student t test and non-parametric tests) - Correlations between subscales RFL and BDI (Pearson Correlation) - Statistical dependence links between pairs of variables (religious belief and group of alcoholics in treatment of belonging, through contingency tables and Chi square test) - Effects of socio-demographic variables and religious beliefs on the average of sub-scales and the items RFL (various applications of Analysis of Variance).

GOALS The goal of this research was to measure the effectiveness of the treatment administered on an outpatient basis, for the rehabilitation of paralyzed limbs. The treatment consisted in the administration of Botulinum Toxin Type B. The efficacy of treatment was about three main aspects: - Function of limbs - Activities carried out with the arts - Participation in social life METHOD This statistical analysis was performed on a sample of 331 patients suffering from stroke, with paralysis outcomes of arts, who have completed rehabilitation on an outpatient basis. To measure the effectiveness of the treatment, a questionnaire was administered, with various items related to the three main aspects mentioned above. Patients were able to record their answers based on a four level scale: 1 = very good; 2 = good; 3 = moderate; 4 = moderate The questionnaire was administered in two different times: after one month and after three months following the start of treatment. The outcome was assessed according to the differences in the scores recorded by the two surveys, in terms of: improvement, no change or deterioration. RESULTS Statistical analysis revealed a significant connection between the number of treatments and changes in the judgments made by patients, concerning the issue of participation in social life after three months, compared to a month (after start of treatment). In addition, it emerged that the judgments of the patients depended significantly by each hospital in which they received assistance. Therefore, it was highlighted the importance of two factors: the number of treatments and the role of the particular clinic where treatments were administered.

A London undergraduate, who was attending a business course, asked me to help him to solve and comment on some exercises of statistical data analysis, while he was preparing for an examination. I thus had the opportunity to apply various methods of quantitative analysis, using the SPSS statistical package and adding to its output my own explanations and charts. The methods included: descriptive statistics, contingency tables, multivariate testing (parametric and non parametric), ANOVA and linear regression. The student has passed the examination with excellent marks.

This is an exercise solved by me for a student of mine. - The client was a luxury brand company, which wanted to launch a direct mail campaign, which targeted some middle-class wealthy people, with some products. - I had to build a model, from the USA public Census data, for forecasting if a person is likely to earn more than 50k USD annually. These people were the target of the above campaign. - I had the availability of a big random sample of more than 32k people, and I built a model able to predict, with a reasonable confidence, if a given individual earned an annual income greater than 50k USD. - I used the software SPSS Modeler 18 and the CRISP-DM methodology. less