CMBS Tenant Normalization Training Dataset

This dataset is derived from the paid dataset https://app.snowflake.com/marketplace/listing/GZT0ZYQ5625/foodtruck-ai-cmbs-schedule-al-xml.

The purpose of this dataset is to provide the fields necessary to train a machine learning model to normalize\classify commercial real estate tenant names.

The schema is devoid of CMBS deal information however contains the property address and tenant information:

PROPERTYNAME: Name of the property, typically for identification.
PROPERTYADDRESS: Full address of the property location.
PROPERTYCITY: City where the property is located.
PROPERTYSTATE: U.S. state of the property's location.
PROPERTYZIP: Zip code for the property's location.
PROPERTYTYPECODE: Code for the property type (e.g., "OF" for Office).
TENANTNAME: Name of the tenant occupying the space.
TENANTSIZE: Size of the tenant's occupied space.
TENANTLEASEEXPIRATION: Expiration date of the tenant's lease.

The full list of 230+ deals used to derive this dataset can be seen here: https://dealcharts.org/capitalmarkets/abs/cmbs/

This dataset contains Schedule AL data from Edgar Form ABS-EE filings for 230+ CMBS deals across 9,500+ filings (monthly). The earliest reported period is 2016-11-07. The dataset is updated daily to capture all future ABS-EE filings as they are published in Edgar.

The Schedule AL is a statement of disclosure of assets and liabilities and is part of Reg AB II Asset-level Requirements Compliance that went into effect on November 23, 2016.

Form ABS-EE is a form that issuers of asset-backed securities use to file exhibits EX-102 (Asset Data File) and EX-103 (Asset Related Document). More information can be found at https://www.sec.gov/oit/announcement/regabii-asset-level-requirements-compliance and https://www.sec.gov/divisions/corpfin/guidance/form-abs-ee-interps.

All Conduits since 2018 are represented in this dataset along with information about their loans and properties. This dataset contains the original XML submission, no processing has been done whatsoever to the data submitted.

搜集汇总

背景与挑战

背景概述

该数据集源自商业地产抵押贷款证券(CMBS)的Schedule AL数据，包含2016年至今230+笔交易的物业地址、租户名称及租赁信息等字段，用于训练租户名称规范化模型。数据每日更新，保持原始XML格式未作处理，符合美国SEC的Reg AB II资产级披露要求。