US TTB COLA Registry Alcohol Product Catalog - Demo

In the US, alcohol product labels must undergo a regulatory approval process facilitated by the Alcohol and Tobacco Tax and Trade Bureau (TTB).

Alcohol producers digitally submit their prospective labels along with supporting documents to the TTB, and these application materials are published to the TTB's Public COLA Registry (COLA refers to a Certificate of Label Approval).

This registry contains over 2MM application records dating back to the late 90's, and the registry adds ~3K new records each week. In addition to the label imagery,
these applications include a form with ~20 structured fields describe the product and the manufacturer.

These records are great representation of alcohol products soon-to-be available in retail locations in the US market.

COLA Cloud scrapes, parses and enriches the TTB's Public COLA Registry. In addition to all the structured fields of each COLA, the COLA Cloud pipeline includes OCR via the best-in-class Google Cloud Vision API on every label image (a $6,000 value over 4M images) to provide label text as a column, and. reads barcodes from label images (~25% of applications contain a unique barcode).

The demo edition of this dataset includes the entire year of 2018. The `colas` table contains ~170K rows. All the features of COLA Cloud are included.

Tables included:
- colas_demo
- cola_images_demo
- cola_image_barcodes_demo

Field Highlights
- Region (enumerated)
- Category (enumerated)
- Business Name
- Business Address
- Brand Name
- Product Name
- Image-extracted barcode type and value (eg UPC)
- Image-extracted full text
- Image-text derived % ABV
- Image-text derived Container Volume and Volume Units

Source:
This data is sourced from the US TTB Public COLA Registry.
https://www.ttbonline.gov/colasonline/publicSearchColasBasic.do

Developer:
COLA Cloud was developed by a Senior Analytics Engineer with experience at Drizly (Uber Eats), and Gopuff, the two leading brands in US alcohol ecommerce.

搜集汇总

背景与挑战

背景概述

该数据集基于美国TTB公共COLA注册表，记录了酒精产品标签的审批申请，包含结构化字段、图像OCR文本和条形码信息。演示版涵盖2018年全年约17万条记录，涉及产品类别、制造商详情及图像提取数据，适用于分析美国酒精产品市场。