FCC's Form 477 • cori.data.fcc

FCC’s Form 477

Simplifying access to data is usually a matter of opinion. What data should we keep? What data should be discarded?

For effective data analysis, it is important to communicate these decisions and provide the full story of the data.

Short introduction

FCC reports on Form 477 about Fixed Broadband deployment. Broadband providers were required to file information:

All facilities-based broadband providers are required to file data with the FCC twice a year (Form 477) on where they offer Internet access service at speeds exceeding 200 kbps in at least one direction.
Fixed providers file lists of census blocks in which they can or do offer service to at least one location, with additional information about the service.

Source: https://www.fcc.gov/general/broadband-deployment-data-fcc-form-477

The census block mentioned are from US census blocks and from the 2010 vintage. Water only census blocks are also removed.

Every row in this dataset matches a specific service provided by an ISP in a census block. An ISP can provides multiple services, such as different technologies or different speeds in a census block.

The time period covered goes back to December 2014. There is an update in June and December for every other year until December 2021.

NOTE A provider is just required to tell if they provide a specific service in a block. A provider is not required to provide how much of that block is covered by their services.

License and attribution

Broadband data from FCC Form 477, and data from the U.S. Census Bureau that are presented on this site are offered free and not subject to copyright restriction.

FCC asks for proper credit to be given:

Source data: FCC Form 477

If you are using these data please follow these simple guidelines.

Source: https://broadband477map.fcc.gov/#/about

Our opinionated version:

From 477 data was released in multiple versions for some but not all releases. We took the decision to only use and ingest the last release of each version. We are assuming here that those versions are more consistent, and so only selecting the last version will help homogenize this dataset.

We also decided that satellite data is less important for most workflows, hence we focused on Fixed without satellite data.

Our opinionated version is perhaps suboptimal for your specific uses. Feel free to look at our implementation making sure that our process is reliable.

We provide all the code used to generated this data set here.

Data dictionary

library(cori.data.fcc)
table_with_options(get_fcc_dictionary("f477"))

You may have notice that we could have been more strict with regard to our data types. For example, an FRN should be 10 numbers (with padded 0), ie CHAR(10), but our experience working with previous ingestions have shown that this was not a consistently enforced rule.

We also decided to provide the data “as is” instead of removing rows with dubious values. Those decisions should be done by the analyst for now.