About » Technical Overview » Post Processing Data
This section describes the process used to transform raw data into the format we display on the National Broadband Map. The intent of this section is to describe the process the Federal Government implements on the data after it is received. In general, we (1) analyze and summarize each submission to inform the government on the quality of the data; and (2) develop statistics for ranking and summarization. We also develop and deliver geographic information layers, which drive the map, and that process is described in the Map Layers Page.
The general process required to analyze and summarize each grantee submission is as follows:
- Download, review and create an assessment package for each grantee's submission. Review includes: record inspection, count and comparison of records to state submitted metadata, and validation that data conforms to geodatabase rules.
- Review the data for format and content, including confirmation that spatial data has a valid topology, and that other types of data have the correct data types and domains.
- Load database tables to enable data quality analysis and map presentation.
- Analyze the following metrics for completeness: provider, technology, households with broadband availability, speed, and community anchor institutions (CAI).
- Construct the spatial data layers for the National Broadband Map.
Rank and Summarize
Random Point Table Analysis - Wireline Statistics
This analysis transforms address and road segment data into consistent counterpart data of blocks in order to calculate broadband availability statistics. This allows us to develop population and housing statistics across the entire dataset. For example, given just individual road segment data, it is not possible for us to determine the number of households on that particular road segment. As a result, it is impossible to find the number of households with broadband availability.
The calculation first estimates the point locations of households in census blocks greater than two square miles given the number of these households and the distribution of roads within that block. Our model follows the basic premise that the distribution of households exactly follows the distribution of roads. We therefore grow households along the distribution of roads.
Once we have this distribution, using geospatial operations, we find the number of modeled household points that are within 500 feet of every broadband availability location (road segment and address point) and divide by the total number of households within that block. This gives us a distribution of households and population within that block. For the purposes of this calculation, when a provider offers service within a block that is less than two square miles, we assume that the provider is offering service throughout the block. Once complete, we then combine the random point results and the blocks less than two square miles results into one table for consistent wireline statistics.
Wireless Block Overlay - Wireless Statistics
This analysis allows us to determine the percentage of land area (and unique identification code) of all census blocks that lie within each wireless shape (e.g., wireless provider coverage area).
Wireless shapes given by broadband providers are free forming amoebic patterns and do not follow established geographic boundaries. For the purposes of calculating broadband availability statistics for a particular geography, the Broadband Map application must know to what extent each and every block falls within, is contained by or is excluded from each and every wireless broadband availability area. To establish this statistic, we overlaid 2010 census block geographies on each wireless shape. The resulting table is a single table with the percentage of land area for each census block within each wireless polygon. We use this percentage as a constant factor in determining the percent population and household availability for that provider, technology and speed.
Incorporating Speed from the Overview Table
This analysis allows us to include maximum advertised download and upload speeds into the census block, wireless overlay, and random point (address and street segment) tables for speed data provided at the county level in the Overview table.
In some instances, grantees submit advertised download and upload speeds at the county level. In these cases, we backfill the maximum advertised speeds from the overview table into the census block, wireless overlay and random point tables for every particular census block, provider, and technology combination where the maximum advertised speed is null. If there is speed data available at both the census block and county levels, we use the data from the census block level due to its higher granularity.
In order to perform this update, we joined the census block, wireless overlay and random point tables with the overview table by county FIPS (using the first five characters from the block FIPS to obtain the County FIPS), as well as FRN, provider name and DBA name. We then updated the maximum advertised speed values in the census block, wireless overlay and random point tables with the speed found in the overview table.