jrjames83

Very nice - can't wait to try. What kind of features and distance measurements do you use for the audio similarity? I've worked on that quite a bit in the past.

I try to avoid doing too much in a single day. 20 mins of the class, followed by an hour of slightly changing the instructions, or imagining, "how could I tweak it to do "this"", usually pays off the best for me, in terms of knowledge retention.

JRE did a good job with that interview. He has some good recent interviews on Jim Rome and Mark Divine (Sealfit). Definitely a good listen once or twice a month to hammer home a proper state of mind.

In case anyone wants to listen to the recent podcast: https://www.macrovoices.com/360-chris-cole-volatility-and-th...

I would recommend working on as many Kaggle competitions as you can. Try your best to understand manipulating data quickly and efficiently (use Python). Using the correct algorithm is typically trivial and a solved problem for most typical business issues (classification of text, imagery or audio data). At first you'll have to look at other people's code and copy/paste/edit/learn/iterate. Also, don't underestimate the complexity of creating good training, validation and potentially test sets.

95% of your time will be spent massaging data. People say it so often, it may sound ridiculous, but I promise you, it's not. This skill is also readily transferrable to other domains.

I would start with classifying text using sklearn and Facebook's Fasttext.

Then try the dogs/cats image classification challenge and get familiar with Keras and its utilities. I created a few hours of content around recognizing Bill Gates or Jeff Bezos, then trying to recognize 2 types of dog breeds. I outline the challenges of creating "good" training and validation sets.

https://www.youtube.com/watch?v=O3hffX-jC98&list=PLImyDqSBQb...

For every hour you spend looking at some equations, you missed an hour expanding your skill-set to manipulate and get data into a format which can be fed into well understood and maintained algorithms. Once you feel like you get produce results, go back and think about the underlying mechanics. My 2 cents from experience.

Best way is to see it in action and experiment. A few videos I made (b and c relate to common queries written for BI or CLTV stuff).

a) row_number over() and partition: in the abstract: https://www.youtube.com/watch?v=-X3eIyZV728

b) applied to customer value analysis: https://www.youtube.com/watch?v=iHxJvF0tZOA

c) applied to time differences (also uses materials from b above) https://www.youtube.com/watch?v=5f8tF4U70Ic

After a while you can write these from scratch and they generally work the 1st or 2nd time, but it takes lots of trial and error at first when setting up the partition and ordering clause within the OVER() expression.

I created a set of videos for Udemy, but they complained about audio quality so I added them to YouTube. I cover a variety of basics (joins) but also get into subqueries, correlated subqueries, window functions and all along use a freely available database. https://www.youtube.com/playlist?list=PLImyDqSBQbdm09n4BhGwK...