insights from locaCon data. To build the ecosystem that will enable individuals and businesses to map their world's data and become more knowledgeable and successful MISSION
SQL API Redis Maps API Torque Mapnik Varnish Caching Varnish Caching CartoDB on Amazon elasLc Recipes CartoDB on Amazon elasLc Recipes PostgreSQL PostGIS CartoDB Server API Rest API Imports API CartoDB Editor Browser
• At the core is a C++ shared library providing algorithms and pa]erns for spaAal data access and visualizaAon. • Combines pixel-‐perfect image output with lightning-‐fast cartographic algorithms • And exposes interfaces in C++, Python, and Node. mapnik
duplicates is a very fast operaAon • Snapping at high tolerances can lead to an output that doesn’t follow the original line parAcularly closely, and can contain some odd effects
any point within a tolerance to be “repeated” • The cost is only the distance calculaAon, so it’s no more expensive than snapping, and the results are much closer to the original line.
uses “well-‐known binary” (WKB) format to represent geometries when requesAng them from PostGIS. • The WKB format in turn specifies 8-‐byte doubles for represenAng verAces. • That can take a lot of space, space that maybe we don’t need to use.
next vertex in a shape usually only takes 1 or 2 bytes, because the points aren’t very far apart • There is a big space savings available in using “delta” coordinates rather than “absolute” coordinates • Niklas Aven was working on “Tiny well-‐known binary”, a representaAon that retains the logical structure of WKB (simple feature geometry types) but packs them into the minimum possible size using delta encoding and variable length integers • All this calculaAng of deltas and minimum sized integers takes a fair amount of CPU Ame, but in our applicaAon the trade-‐off is worth it