Data Science in Application Performance Management

In this talk, we’ll use real world application data to illustrate key concepts in visualizing and analyzing timeseries performance metrics in modern web applications. We’ll scrutinize heatmaps, frequency charts, histograms, rug plots, box plots, scatter plots and more, breaking down their strengths and weaknesses and assessing their cost/benefit tradeoffs.

We’ll also cover statistical properties of timeseries data, comparing various summary metrics and breaking down the different kinds of distributions present in application performance metrics.

All data and analyses will be made publicly available.