The analyses developed used ABEXUS’ proprietary data engine. This engine is an ensemble modeling framework in which multiple analysis blocks are executed within a state-of-the art data representation environment. The model ensemble blocks which are then pitted against each other on the basis of several custom criteria, including: reliability metrics, prior and posterior credibility rankings, etc. Computed models that have favorable scores in these criteria are then selected and further refined via ancillary analyses deployed based upon a semantic AI symbolic analysis, which essentially codifies the logic behind the underlying questions answered by the overall ensemble. A selection of the results provided by the data engine are then incorporated into the analysis as either visualizations or results.
At the county level, the engine was also able to identify differences in Puerto Rican communities throughout the US. Moreover, the system identified the three main states with high concentrations of Puerto Rican population and their main socio-economic profile. This analysis was also carried out for the three counties with the highest Puerto Rican population in the US.
To improve the accuracy of our analytics, we analyzed a total of 3,142 counties, 73,056 census tracks, and 1,738,320 Puerto Rican households. To further improve our insights, we examined the consumption patterns of such groups using data from Puerto Rican goods sold in the U.S. by Brands of Puerto Rico. Data and illustrations are derived from AX’s data engine, which includes an aggregation of 3,142 counties, 73,056 census tracks, and 1,738,320 Puerto Rican households, plus, proprietary data from Brands of Puerto Rico customer transactions and analytics.