Last November, CIS Consulting, a company of Spaulding Ridge hosted first face-to-face edition of Data & Wine in Madrid. When purchasing a bottle of wine, we often rely on the smell and flavour descriptions, winery reputation and price. Our analytics team sought to prove that you don’t need to be a sommelier or a connoisseur to make the best choice, so long as it’s based on trusted data!
For this event, we used a dataset containing over 130,000 wine reviews, wherein each entry specified the origin, designation, price, taster’s description, and rating. We began by uploading the dataset to Google BigQuery, then conducted a preliminary analysis to check for any data quality issues and proceeded to model the data on Looker. By extracting insights from both quantitative and qualitative data, we wanted to determine what makes a wine great, and whether there is a clear correlation between wine’s origin, price, description, and points.
The dataset contained reviews from 46 countries, with 5.48% corresponding to Spain. After visualizing the distribution of wine price and rankings, we could confirm that, on average, as the age and price increase, so do the points – with a mid-priced vintage bottle rated 88. However, seeing that the most expensive wines were not best rated while some cheaper wines received full marks, and that price grew exponentially with review score, we had to consider other properties and descriptors.
Looking at the price, score and number of reviews correlation, we noted that the spending per point increases significantly among top-rated wines, which implies that tasters may be biased. We then introduced meteorological data, calculated best value by country and wine variety, and conducted sentiment analysis to provide contextual insights. By detecting positive, neutral, and negative sentiments, we could determine how tasters feel about different wines and what makes a particular product popular. The visibility Looker enables is instrumental in identifying which wines are likely to sell best, and we can draw conclusions backed by accurate data.
To learn more about the possibilities Looker affords, we invite you to read the Vivino case study. With over 12 million wines and 44 million users, Vivino is the world’s largest online wine marketplace and most downloaded wine app. The company is using Looker to enable organization-wide access to reliable insights, drive sales, assess customer satisfaction, and fine-tune retention strategies.
By the end of our demo, the attendants gained a better understanding of how Looker can be used for interactive, collaborative visualisations and turning valuable insights into data-driven decisions.