Wine and Data Science

[GregT]

Hi all.

CellarTracker is copyrighted. We do not have an API. We have actually had a lot of problems with academics attempting to harvest large amounts of data for their own analysis and which they later loaded into the “public domain” against our terms of service. This was the worst example: SNAP: Web data: CellarTracker reviews

It actually took multiple cease and desists to get them to comply, and a few years later they even “restored” it one more time for good measure.

Did they say what their purpose was? If it’s just to get ratings, they can already do that just by looking at CT. A few stores have asked me how to find out about ratings. The owners don’t know anything and in one case, for religious reasons, doesn’t even taste wine. I’ve steered them to CT.

Hope that’s OK.

Too bad these guys ended up with such a pointless paper. They started with an assumption and created a hypothesis that they thought would support that assumption:

"The very act of consuming products will cause users’ tastes to change and evolve. Developing new models that take into account this novel viewpoint of user evolution is one of our main contributions.We model such ‘personal development’ through the lens of user experience, or expertise. Starting with a simple definition, experience is some quality that users gain over time, as they consume, rate, and review additional products. The underlying hypothesis that we aim to model is that users with similar levels of experience will rate products in similar ways. . . "

Then they concluded by re-stating part of their assumption and tacking on the other part that they couldn’t let go, and burying what was probably the essential support for the assumption.

CONCLUSION

Users’ tastes and preferences change and evolve over time. Shifting trends in the community, the arrival of new products, and even changes in users’ social networks may influence their rating behavior. At the same time, users’ tastes may change simply through the act of consuming additional products, as they gain knowledge and experience.

[Eric_LeVine]

I’ve steered them to CT.

Hope that’s OK.

It’s totally OK. Our data is open for consumers and trade people to browse on our site. They just can’t machine harvest it for their own purposes.

[GregT]

They aren’t. Just look it up themselves and they also put up notes in the aisles referring customers to it.

[Roald_Schuring]

fascinating thread for sure. professionally, i think about these things all the time. Roald, i’d love to chat directly if you’re up for it.

Would love to chat! Let me send you an IM.

Roald, I think they are trying to do something different than you are, but here is another team of data scientists working to apply their craft to wine:
https://www.globalwinescore.com/
Their approach is more about aggregating and then normalizing critics scores (not words), and determining the level of consistency for a specific wine.

I wasn’t aware of this project. Very interesting, thanks for sharing!


Thanks everyone for the lively discussion so far. I’ve been struggling to get good feedback on my work so far, and this forum has been really proactive. I’m starting to land in a place where I think doing analysis on the text contained within individual wine reviews is probably a bit too optimistic. I do think that there is great value in the information once these reviews start to get aggregated at some level (whether it be geographic or varietal-based) and reach a certain sample size.

Another takeaway is that I may be attributing too much importance to ‘expert’ opinions. A large number of amateur reviews might contain as much - if not more - valuable information than a smaller number of ‘expert’ reviews.

Please keep your ideas & thoughts coming…!