PREDIK Data-Driven I Enriched Address Data for Commercial & Industrial Locations I +50 million US addresses

Empower your machine learning models with a curated dataset containing more than 50 million commercial and industrial addresses in the USA. Featured attributes of the data - Visualization data: Number of devices detected, time spent, income level, possible workers, trucks, among others. - Structure data: Built areas, parking lots areas, POIs. - Activity data: Commercial/Industrial activities at each address. - Device count data: Device count for every commercial and industrial location in the US. - Visitation data: Visitation data for every address in the US. This dataset has proven successful in enriching machine learning models for defining POI/land parcel activity, sales prediction, site selection, red flags systems, procurement, and commercial/industrial activity measurement, among others. The dataset includes 54 attributes for each of the 50 million addresses/parcels. Some of them are: - Lat and long. - Ave. sq ft. - Ave. sq ft. per POI within the parcel. - Significant Dwell time of the visitors. - Ave. visitor stay time. - POIs in facility. - Trucks. - Possible workers. - Income of the visitors (percentile). How have our clients used this dataset? Cold Storage Company: - Data requirement: Our client needed more data about company locations that provided cold storage solutions, as part of their sales & marketing strategy. - Solution: They ingested the US address dataset in their ML model to enhance the process of identifying all facilities within the US with cold storage solutions (From Specialized Cold Storage facilities to Distribution Centers that handle refrigerated products). Tax Advisory firm: - Data requirement: Our client needed information regarding commercial activity to ingest in a machine learning model that generates Red Flags for potential tax evasion. - Solution: Our dataset provided detailed information with several data attributes for each of their target addresses. Also, our client received crucial insight for their investigation process, such as visitation rates, people working, proximity to other commercial locations and visitor's income level.

搜集汇总

背景与挑战

背景概述

该数据集包含美国5000多万个商业和工业地址的54项属性信息，涵盖空间特征、访客行为、经营活动等维度，已成功应用于销售预测、选址分析等机器学习场景。数据集通过两个典型案例展示了其在冷存储公司选址和税务咨询公司反逃税模型中的实际应用价值。