My Favorite Data Science/Machine Learning/Statistics Resources

The Good, The Bad, The Ugly

 Posted on March 9, 2020  |  10 minutes  |  1944 words  |  Michael Johnson
People often ask me, “what resources do you use to keep on top of machine learning”? I spent some time curating some of my favorite resources and I figured this would be a fine place to share them. But first, there is some bad news. The best resource I know of for staying abreast of developments in the field is Twitter. I know, this is not what you (or I) wanted to hear, but time and time again Twitter has shown me new papers or new findings that have had significant impact on what my team does. [Read More]
forecasting  Deep Learning  machine learning  Artificial Intelligence 

Distributed Graph Representations Using the Mueller Report

Node2Vec in 14 lines of code

 Posted on January 25, 2020  |  11 minutes  |  2219 words  |  Michael Johnson
Note: This is part 2 of my series on the Mueller Report. Please take a look at Part 1 if you are interested in how I built the graph. One thing that would be useful when navigating a document (or set of documents) like the Mueller Report is the ability to find things that are ‘like’ other things. For example, if you are trying to follow the thread of a story through the document, you might want to find all the paragraphs that are about similar things to the paragraph you are interested in. [Read More]
nlp  spacy  graph  trump  mueller  python  visualization  pyvis  networkx 

Graph Visualization of The Mueller Report With SpaCy and PyVis

Building a contextual paragraph recomendation engine

 Posted on June 19, 2019  |  12 minutes  |  2350 words  |  Michael Johnson
One of the most interesting talks I heard at Strata in San Francisco this year was “Towards deep and representation learning for talent search at LinkedIn”. In the talk, Gungor explained how he took advantage of LinkedIn’s economic graph to build a hyper-personalized search engine. Ever since then I’ve had graphs firmly planted in my mind. Not these graphs: Graph More like these: Network Specifically, I’ve been trying to understand how graph network techniques can be applied in various domains, including natural language processing. [Read More]
nlp  spacy  graph  trump  mueller  python  visualization  pyvis  networkx 

Monte Carlo Simulation For Childrens Stories

Using web scraping, SpaCy, and monte carlo simulation to check children story logic

 Posted on January 25, 2019  |  13 minutes  |  2621 words  |  Michael Johnson
Bedtime has historically been a battle for our family. Kids have this impressive ability to fall asleep when you want them awake and vigorously stay awake when you want them to sleep. To combat the insanity, we read a few books every night before bed. There is a great series of books for young children called “The Magic Tree House”. The series follows two children (Jack and Annie) who transport to other times and places for some mystery or adventure. [Read More]
web scraping  spacy  nlp  python  simulation  pandas  family  monte carlo 

How Many Boys is Too Many?

  Posted on January 19, 2019  |  10 minutes  |  2026 words  |  Michael Johnson
Anyone who has multiple boys in a row at some point asks themselves a question: did I just get ‘lucky’ or are we only able to have boys? As you can see, we certainly asked that question… When we had our third boy in a row I started to wonder: how many children do you have to have of a given sex before you can statistically conclude that there is something other than random chance influencing to the outcome? [Read More]
hypothesis testing  R  simulation  probability  tidy  family 

Deep Autoencoder Neural Networks for Maximal Christmas Decoration Enjoyment

  Posted on November 21, 2018  |  8 minutes  |  1528 words  |  Michael C Johnson
My wife loves decorations. From Valentines Day to Easter to Thanksgiving, our house is adorned with interesting festive items. Her favorite season by far is Christmas. Every December we drive around trying to find the best Christmas lights. Inevitably, there are a few houses with lights that blink with the music, something I’ve always been fascinated with. Two years ago I embarked on a mission to build my own Christmas Light Show, specifically this is what I had in mind: [Read More]
Deep Learning  Raspberry Pi  python  Keras  tensorflow  Christmas  FFT  Autoencoder