Deakin University MIS772: Predictive Analytics - Wine Rating Report

This report presents a predictive analytics case study aimed at assisting Australian Wine Importers in making informed decisions regarding wine selection. The study focuses on two key questions: identifying wines similar to a new introduction and estimating the rating of the new wine in the Australian market. The analysis utilizes a dataset of 1300 wine entries, encompassing variables such as wine title, country, province, region, variety, winery, description, designation, and price. The methodology includes cluster analysis and text mining to compare wines based on origin and rating points, alongside segmentation analysis to examine relationships between wine characteristics and ratings. The results indicate that new wines from the USA, France, and Italy have the highest rating points, suggesting that these countries are prime sources for importation. The report also highlights that higher-priced wines tend to receive higher ratings. Overall, the analysis provides valuable insights for Australian Wine Importers to optimize their wine selection strategies based on data-driven predictions.

PREDICTIVE ANALYTICS
1
PREDICTIVE ANALYTICS
Assignment A2 / Workshops M1-M3: RM
Name of Student:
Name of Institution:

PREDICTIVE ANALYTICS
2
Problem
This is a report about business predictive analysis case study. The problem that this report is aiming to solve is providing a clear
prediction that can be used to guide the selling of wines in Australia. The purpose of this report is to help Australian Wine Importers
to answer two critical question: The group of wines the new wine is most similar to and why and knowing the estimated rating of
newly introduced wine to the Australian market. Therefore, the aim of this report is to explore the wine data and visualize the various
rating points of the wines. Therefore, this report is targeting to improve the decision making process of Australian Wine Importers on
the right kind of wines to import.
Data Preparation
The data that has been used for analysis consists of several variables about the wines. The total number of entries or observations of
the data set is 1300 (the total number of columns). The variables of the data set included: Wine “title” (name + vintage), Country,
Province and Region, Variety and Winery, Description and Designation and Price (US$).
Executive Report
In order to achieve the objective of this study, the following analyses have been conducted: Cluster analysis of the wines text,
conducting segmentation analysis to determine the relationship between the wines and visualize the results. Clearly, the data consists
of both text and numerical data. The analysis involved both texts and numeral analysis. The text mining techniques were meant to
discover and compare the wine brands, the country of origin or region and the comments of rating that a particular wine received

PREDICTIVE ANALYTICS
3
(Anand, et al., 2015). Text mining techniques that have been used include cluster analysis and word cloud analysis. The analyses have
been to compare the new wine and the other wines as well as to compare the wines based on the regions or the countries of origin.
Segmentation, which is a form of text analysis, has also been done (Gunasekaran, et al., 2016). Segmentation has been used to
separate the compare the wines based on the countries of origin and the prices and points. Visualization of the data sets have been
done using several tools such as the histograms (Kopcso, et al., 2018).
The classification results demonstrates that there is a significant difference between the new wine and the other wines. The results
show that new wines from USA, France and Italy have the highest rating points. Therefore, it is statistically sufficient to conclude that
it would be profitable to import new wines from the three countries (Ancel, et al., 2015). The other result demonstrate that the prices
of wines from USA and France are the highest. The results suggest that people love wine with higher prices since the wines from USA
have the highest rating points and the highest prices as well (Hazen, et al., 2014). Therefore, based on the analyses, it is clear that the
new wine is similarly to wines from three countries: USA, France and Italy. Based on the reviews, Australian Wine Importers should
put more focus on wines from these three countries (Tsaih, et al., 2018).

PREDICTIVE ANALYTICS
4
References
Anand, A., Kant, R., Patel, D. P. & Signh, M. D., 2015. Knowledge Management Implementation: A Predictive Model Using an
Analytical Hierarchical Process. Journal of the Knowledge Economy, 6(1), pp. 2-5.
Ancel, et al., 2015. Predictive safety analytics: inferring aviation accident shaping factors and causation. Journal of Risk Research,
18(4), pp. 1-7.
Gunasekaran, et al., 2016. Big data and predictive analytics applications in supply chain management. Journal of Computers &
Industrial Engineering, 101(2), pp. 1-9.
Hazen, b. T. et al., 2014. Data quality for data science, predictive analytics, and big data in supply chain management: An introduction
to the problem and suggestions for research and applications. International Journal of Production Economics, 154(08), pp. 2-6.
Kopcso, David, Bachamanova & Dessiaslava, 2018. Case—Managing Staffing Inefficiencies Using Analytics (B): Business Value in
Predictive and Prescriptive Analytics Models. Journal of INFORMS Transactions on Education, 19(01), pp. 3-10.
Tsaih, et al., 2018. The use of big data analytics to predict the foreign exchange rate based on public media: A machine-learning
experiment. Journal of IT Professional, 20(02), pp. 2-8.

PREDICTIVE ANALYTICS
5