When using Amazon Kinesis to process a large volume of incoming lat/long data points, what's the best way to reverse geocode these at scale?
Using Kinesis Analytics I can use reference data stored in S3, as documented here: http://docs.aws.amazon.com/kinesisanalytics/latest/dev/app-add-reference-data.html
This could contain country & city lat/lng data e.g. from Geonames.
Would it then be possible to run an Analytics query with a haversine formulae to calculate the nearest closest city in my reference table to my lat/lng coordinate data? I could then create an output stream with the enriched reverse geocoded data.
It looks like the SQL engine is missing COS/RADIANS functions however.
Any help would be greatly appreciated.