Category : Engineering

Better Whisky Drinking Through Data Science

Nanigans is known for its hardworking, fun-loving culture so our love of Whisky Data Science should come as no surprise. This post is a discussion and visualization of the tasting profiles of Scotch whisky from 86 distilleries. We use a form of dimensionality reduction to create a tasting map that shows similarity between different whiskies. As a big fan of the amazingly smoky Laphroaig from Islay, I’d also like to identify some good candidates for my next Scotch to try. […]

How to Boost Performance with In-Memory, Multi-Dimensional Analytics

Nanigans’ ad optimization platform, Ad Engine, leverages data to provide the best ROI (return on ad spend investment) for our customers. Making this data easily accessible in a myriad of dimensions is pivotal to analysis. Scarecrow is the backend engine that replaces our original Placement Analysis utility focusing on high performance despite large data size (a million+ ad placements) with the ability to incorporate multiple, custom pivots across an ad’s creative, audience, and performance over time. Background The original version […]


How to Create a Table That Can Display One Million Records

If you were to create a PHP script that creates a plain HTML table with one million rows, you’ll immediately run into a few problems trying to load it in a browser: It requires downloading hundreds of megabytes (depending on row size) The browser will create millions and millions of DOM elements and run out of memory As a result, it will freeze and crash. Developers have known that this is a foolish thing to do for a long time.  […]