Visualizing large data sets
I'm going to showcase how to visualize large data sets, i.e. that have thousands to millions of data points.
This goes beyond standard techniques like bar plot etc, and requires using tools like d3, Processing, ggplot2, circos and more.
I will demostrate working samples, that I have created, using open source tools.
Folks will gain an understanding of concepts, techniques and tools to create large data visualizations.
One of the best ways to explore and understand a dataset is with a visual.
We basically depict the data into a visual space and let our brain explore the patterns.
Later, when we want to present the findings to others, visualizations are essential for efficiently communicating our findings.
However, this is challenging with large data sets. Whatever it is, a good data visualization, big or small, aesthetically appealing or not, helps us see what the data has to say. And if a visual cannot give that insight, then it’s not that useful.
Rather than going through a exhaustive list of visualization tools, that may be searched on the Internet, I'm going to show you some working demos that I have created on large data sets, on products that I have worked on in Yahoo and outside.
So the ideas I'm going to share in this tech talk are focused on both usefulness & aesthetics. And for someone who's fond of both art & science, this is fantastic area, this is a great area as we are limited by only what we can imagine.
The only prereq for this tech talk is your level of curiosity, you dont need to be a programmer to take advantage of these visuals I'm going to show. This tech talk, thus, is less academic and more targeted on creating some good visualizations that can simplify analyzing complex data.
More than 15 years in the software industry on web, mobile, analytics, video, edge and visualizations.
Lived and worked in US, Spain, UK, India.
BE from BIT Ranchi India, MS from NEU Boston MA USA, EPBM from IIM Calcutta.
Product Head of performance & analytics at Yahoo.