Wine analysis
Description
That was our mission. With site scraping, the site should be informed about its information and the advice should be useful to the buyer’s business. This should have affected the growth of the site. The customer also needed a program that would predict the probability of buying this wine at a reasonable price when the program entered certain elements. That is, the developer did not want to use many experts and asked us to create a program that would act as an expert. 🤠
Data visualization
What are the most popular wino colors on the market?
The 30 most expensive wine names.
The most widely used grape variety.
Countries that produce the most wine.
Which years produced the most wine?
The effect of the rating on the price.
Machine learning
We mostly used the sklearn library in the ML part. For prediction, we used Linear Regression. We used Mean absolute error (MAE) for testing.
Get Salohiddin Qayumov’s stories in your inbox
Join Medium for free to get updates from this writer.
Since the scraped data contained many NaN values, it was an error to fill it with a medium value, so we removed the row with NaN values. For this reason, 1600 data with 16000 lines remained.
Summary:
Wines made from grapes with a similar rating should be increased. The wines of the countries that sell the most should be increased. in general, higher rated wines should be increased. You need to sell more wine but spend less money on it. mainly need to sell 750ml wines, and need to buy a server to increase site speed. The apparatus should be increased and the number of wines sold on the site should also be displayed.
