Using Twitter's Breakout Detection Package
summary
Twitter's breakout detection package finds abrupt sustained level shifts in univariate series, originally in R then ported to Python, and sees work on S&P 500 Yahoo and Netflix examples. It computes a robust energy‑distance (median based) across splits, returns the split with maximal score, and offers EDM and EDM multi modes with tunable beta and degree penalization. Requires scaling data to 0–1, tolerates spiky noise better than mean‑based methods, and scales to hundreds of thousands of series as used at C2FO.
AI-written summary · about
chapters (13)
- 0:00Breakout detection package intro
- 1:04What is a level shift
- 2:31Why level shifts matter
- 3:53Level shifts in stocks
- 5:20Package performance
- 6:02Practical uses of shift detection
- 8:08Exploratory analysis
- 9:15Installing the package
- 11:03Basic usage guide
- 12:30Detecting multiple breakouts
- 13:35Adjusting beta sensitivity
- 15:08Algorithm math foundation
- 21:43Q&A
related talks
29mInsights into Social Media Data using Entropy Theory
27mExploring Science on Twitter with IPython Notebook and Python Pandas
24mPython, Pandas and other animals to survive in the data jungle
27mTweaking the rise and fall of empires and economies
45mInteractive Analysis of (Large) Financial Data Sets
35mUsing Scattertext and the Python NLP Ecosystem for Text Visualization
41mFootball's data analytics in python
27mTime series feature extraction with tsfresh - “get rich or die..