Re: Difference btw nanocubes and technology used for google maps
A nanocube is an efficient in-memory encoding for a rollup data cube (think of a data structure or a database index). The benefit of encoding a dataset in such a format is that it enables solving multi-dimensional aggregate queries at a really fast pace. For example, in a crime data set we could ask things like: "what is the hourly time-series of crimes in this arbitrary region I just brushed?" or "what is the heatmap of crimes on weekends vs. weekdays?" and get the answers right away. Solving such queries quickly is specially important if someone is interested in providing an exploratory visualization experience to a user/analyst. So, to answer your question more directly, think of nanocubes as a technology to overlay data (heatmaps, timeseries, barcharts) on a Google Maps-like front-end. In many interesting cases, the nanocube encoding is efficient enough to store hundreds of millions to billions of records in commodity RAM space and allow for a fluid visual interactive experiences (see our technical for more details http://nanocubes.net/assets/pdf/nanocubes_paper.pdf).