Implementing a Real-Time Multi- dimensional Dashboard
The Problem Statement
An analytics dashboard must be capable enough to highlight to its users areas needing their attention. This Rolex Replica needs to be done in real time and displayed within acceptable display time lag to the user. Any screen must be displayed within industry standard time of 3 sec’s. You would need to grill your data along a large number of dimensions. You can make your data talk if you can grill it along the right dimensions (time, location, site, access device ).
[Fig 1-Multi dimensional drill down dashboard]
The screenshot shows SigView Dashboard, it shows the drilldown for Coca Cola Ads in USA with people viewing the ad from a Laptop using OS X.
The transaction and log data being generated in a current systems in humongous. You might want to grill your data along 100’s of dimensions. But for a dashboard to support 5 dimensional drill down along 100 dimensions mean (100C5 = (100*99*98*97*96*95)/(5*4*3*2)) possible rollups. This means that it’s impossible to compute all pre roll ups. omega seamaster 300m replica
To solve this problem, SigView uses a hybrid approach. Storing data in columnar format, computing pre roll ups on all dimensions but only up to a certain point and then resorting to use the raw data when more you try to grill down on more dimensions. Columnar storage along with low cardinality allows for very high compression. The dashboard considers time as a filter, with hour, day and week level of granularity. The dashboard is powered by Apache Spark and can display any result in less than 2 sec’s for up to 60 GB’s of data processed per minute.