There is a severe lack of public data on businesses in the US. While there are some sources these suffer from various issues rendering them useless for a more nuanced understanding of business dynamics within the economy.

• Overly aggregated: For top level indicators of business health—namely the U.S. Census Bureau, Business Dynamic Statistics (http://www.census.gov/ces/dataproducts/bds/)—these datasets are aggregated at such a high level that they are rendered useless for anything other than a high level overview of state and nationwide business health.
• Not accessible/consumer facing: some datasets exist, such as Dunn & Bradstreets company level data used for running credit checks or Yelp's massive database of local businesses, but these datasets are either hidden behind expensive paywalls or protected by copyrights that render them inaccessible to the average data consumer.
• Incomplete and heavily fragmented: Many of the richer open datasets that do exist, such as San Francisco's list of locally registered businesses (https://data.sfgov.org/Economy-and-Community/Registered-Businesses-sorted-by-End-Date/vupz-6rcv) are incomplete and lack sufficient means e.g. IDs, by which to joi**n them with other datasets. The lack of other datasets at the local level furether hinder enriching this data.
• Untimely: Most business datasets mentioned above are not up-to-date in that they are either based on the last time a business submitted information to the local government, or the most recent census that was conducted. As such for temporally sensitive decisions, the data is lacking in sufficient granularity to be useful.