Analytics

Purpose

The Analytics section describes our attempt to make some sense of all the numbers we assiduously collected since 2009. What insights can we get from our observations, and how do those insights mold our evolving strategies in the vineyard and the cellar?

This analysis has been made possible using Anthropic’s Claude-Code artificial intelligence tool. On my own, I would never have been able to efficiently sift through our 800 megabytes of data, select the appropriate mathematical models, write the complex programs to analyze the data, draw conclusions, and present them coherently. At the end of each page is a list of references to academic literature. I have not read them all. My job was to make choices on the direction of the analysis along the way – so, if you disagree with the results or insights, I am the one to blame.

At the end of each section is a link to a PDF file – that is a better format to read than this HTML page


Current Sections

This section is evolving rapidly, and much of it is in draft form. Currently, it contains the following sections:

  • Spectra and Color Evolution: Filtering spectral data from fermentation and cellar samples and computing colorimetric parameters allows us to track the changing color of wine from harvest through bottling. 

  • Estimating Phenolics: We moved from using an external service provider to internal estimation of phenolic parameters. This section shows how and why we did this.

  • Modeling Berry Maturation: We model how berries develop after veraison. The important variables are time & temperature, which affect acidity, sugar content, and color. The purpose of the model is to project berry characteristics weeks ahead of harvest, so that the decision on when to pick is well-founded. 

  • Weather and Vineyard Activities: In-depth analysis of data from our own weather stations revealed errors and omissions. Data cleansing and augmentation are critical. Once we had clean data, we analyzed the timing of vine development stages (flowering, budbreak, and veraison) and related activities (pruning, spraying, and harvesting). Most importantly, we use the Berry Maturation model to decide when to pick..

  • Modeling the Fermentation: We built mathematical models describing the changes of chemical compounds as must is transformed into young wine during the fermentation process



more to come …