ClearScore Dataset | Individual Tickers UK Consumer Transaction Data | 1.4m users.

This is our aggregated version of our Exact.One dataset. You can aggregate the transactions using different variables according to your needs: • Per merchant/ticker. • Per category/industry. • You can choose if you would like to see daily/weekly/monthly/quarterly aggregated data. What makes ClearScore's Exact.One dataset unique? • ClearScore provides consumer debit and credit card data at a transaction level. • The data is made available directly from open banking connections that users have with the ClearScore App. • A large active panel of 1.4m users ever connected and 1.8m accounts • Historic view of data spanning 5+ years. • Native categorisation methodology curated over the last 10 years. • Coverage of over 250 million transactions annually mapped to 330+ publicly listed companies. What is the panel size? • More than 1.4m users who have connected to share their data via Open Banking. 1.8m total accounts ever connected in the panel. How do I receive the data? • S3 bucket transfer (Preferred) • SFTP • Snowflake What’s the structure of the data? • Raw transactions (row-level data) or Aggregated data *Data dictionary available upon request. What is the quality of the panel? • The users who we have acquired have a re-auth rate of ~65%. • The coverage is 250 million transactions accounting for £6 billion in spend (Jan - Dec 2021 example) • Coverage of over 1400 merchants mapping to 330 tickers, 130 of which are UK publicly listed Exact.One is built on an industry-leading transaction categorisation service: • Our categorisation service is a rules based deterministic model which favours accuracy over coverage • We focus on accurately categorising spend at merchants, as well as spend pertaining to credit risk (e.g. income, gambling, benefits, financial institutions, cash withdrawals, and debt management services) • Clients can request improvements to the model, and these can easily be implemented by adding new rules or adapting existing rules Purpose tagging: We classify transactions utilising 286 purpose tags which are rolled up to higher level tags (e.g., childcare benefits > benefits > income). Merchant tagging: We tag 1.4k merchants in our model with updates applied each month. Version control: Strong version control allows us to improve our categorisation each month, whilst not breaking models. Unrivalled foundation: Engine trained on the richest data bank in the UK, with >1bn transactions from 60+ financial institutions

本产品为我们针对Exact.One数据集的聚合版本。您可根据需求通过不同维度对交易数据进行聚合：
• 按商户/股票代码（ticker）维度
• 按品类/行业维度
• 可自主选择查看日度、周度、月度或季度聚合后的数据

ClearScore旗下的Exact.One数据集有何独特之处？
• ClearScore可提供消费者借记卡与信用卡的交易级明细数据
• 数据直接源自用户通过ClearScore应用程序建立的开放银行（Open Banking）连接
• 拥有超大规模活跃用户面板：累计有140万用户完成数据授权连接，关联账户总数达180万个
• 可提供跨度超5年的历史数据
• 自研的交易分类体系历经十余年打磨完善
• 年覆盖交易规模超2.5亿笔，关联330余家公开上市企业

样本池规模如何？
• 累计有超140万用户通过开放银行（Open Banking）授权共享数据，样本池内关联账户总数达180万个

如何获取该数据集？
• S3存储桶传输（首选方式）
• SFTP传输
• Snowflake数据仓库对接

数据集结构如何？
• 原始交易数据（行级明细数据）或聚合后的数据
*数据字典可按需索取

样本池数据质量如何？
• 已获取的用户数据重新授权率约为65%
• 2021年全年数据示例显示，该数据集覆盖超2.5亿笔交易，交易总规模达60亿英镑
• 覆盖超1400家商户，关联330余个股票代码（ticker），其中130家为英国公开上市企业

Exact.One数据集基于行业领先的交易分类服务构建：
• 该分类服务采用基于规则的确定性模型，优先保障分类准确性而非覆盖范围
• 我们专注于精准分类商户消费，以及与信用风险相关的交易支出（例如收入、博彩、福利、金融机构服务、现金取现及债务管理服务等场景）
• 客户可提出模型优化需求，通过新增规则或调整现有规则即可快速落地实现

交易用途标注：我们通过286个用途标签对交易进行分类，标签可向上聚合为更高层级的类别（例如育儿福利 > 福利 > 收入）。
商户标注：我们的模型已覆盖1400余家商户，每月更新标注信息。
版本控制：完善的版本管控机制支持我们每月迭代优化分类体系，同时不会对现有模型造成破坏。
无可比拟的底层基础：模型基于英国规模最大的交易数据库训练而成，累计处理来自60余家金融机构的超10亿笔交易。

搜集汇总

背景与挑战

背景概述

该数据集是ClearScore Exact.One的聚合版本，包含通过开放银行连接的140万英国用户的信用卡和借记卡交易数据，历史跨度超过5年，每年覆盖超过2.5亿笔交易。数据提供原始或聚合格式，采用基于规则的分类模型，映射到330多家上市公司，并通过S3、SFTP或Snowflake方式传输。