Skip to main content

Timeseries Data Superpowers: Intuitive Understanding of FIR Filtering and Fourier Transforms

William Cox (Distil Networks)
Computational Thinking
Portland 255
Average rating: ***..
(3.67, 9 ratings)
Slides:   1-PPTX 

So you’ve collected some great timeseries data: your heart rate while watching The Hobbit, the amount of cheese you eat per day, the price of AAPL stock, or flowrate through your CPU water cooler. Now what? It’s so messy it’s hard to tell what to do. Woulnd’t it be great if you could look at this data in a new dimension? The fast fourier transform can do just that. Too bad it’s so confusing.

The fourier transform, popularized in the fast fourier transform (FFT) algorithm, can be a mysterious subject to many programmers. It plays a major part in many aspects of our digital world – filtering, analyzing data and various algorithms (looking at you, Autotune). Related to this topic is digital filtering – removing high frequency or low frequency parts of your signal. In this talk we’ll explore how these two things are related and develop an intuitive understanding of the algorithms. We’ll be using the Python Scipy package to explore various filters and time-to-frequency conversion. Once we’re done you’ll have to tools to look at the world of data in a new, superpowered, way.

William Cox

Distil Networks

When William isn’t busy being a husband and father, he is an electrical engineer specializing in signal processing and machine learning. He’s worked on underwater robots, radar detection, algorithmic forex, and torpedo tracking. He tweets @gallamine and blogs at http://gallamine.com

Comments on this page are now closed.

Comments

05/21/2014 2:00pm PDT

If anyone has specific data they want me to show an example of, let me know. Also I’ve uploaded some animated images showing some of the stuff I’ll be working through. http://i.imgur.com/IU4pCmA.gif