amunategui/Hopkins_BuildingDataProducts
Public repository.
Discovered public repositories for amunategui in the GitHub catalog.
Public repository.
Walkthrough code for http://amunategui.github.io/yelp-cross-country-trip/
PDF version of full R walkthroughs: R - General Purpose Machine Learning Scripts copy.pdf
Companion code for walkthrough: http://amunategui.github.io/stringdist/
Companion code for http://amunategui.github.io/supervised-summarizer/
Code to walkthrough - http://amunategui.github.io/google-trends-walkthrough/
See how easy it is to download, visualize, manipulate stock market data with the <b>Quantmod</b> library and use all of it to build a complex trading model.
Public repository.
A JavaScript PDF generation library for Node and the browser
Reducing High Dimensional Data with Principle Component Analysis (PCA)
A python library for implementing a recommender system
Public repository.
Public repository.
Public repository.
Public repository.
Public repository.
Companion code for YouTube video: https://www.youtube.com/watch?v=Og7CGAfSr_Y&feature=youtu.be
Companion code for YouTube video: https://www.youtube.com/watch?v=zTlbMHw9CeY&feature=youtu.be
Coursera Practical Machine Learning Rep.
Public repository.
Ensemble/Blender example in R using Caret (companion code for YouTube video: https://www.youtube.com/watch?v=k7sTiTWWCXM)
My take on the fastest possible translator for CSV to VW (vowpal wabbit) files using R. This is a work in progress
Shows how to transform categorical and textual data into dummy variables using Caret's dummyVar function. Code for YouTube presention: https://www.youtube.com/watch?v=7rgzCjrIA-o
Using the function read.table() to break file into chunks to loop and process them. This allows processing files of any size beyond what the machine's RAM can handle. Companion code for youtube: https://www.youtube.com/watch?v=Z5rMrI1e4kM
Walk-through of sparse matrices in R and basic use of them in GLMNET. Companion code for Youtube talk: https://www.youtube.com/watch?v=Ysh2gs8VKvQ
Youtube companion (https://www.youtube.com/watch?v=1Mt7EuVJf1A&feature=youtu.be) - Brief introduction to the SMOTE R package to super-sample/ over-sample imbalanced data sets. SMOTE will use bootstrapping and k nearest neighbor to synthetically create additional observations. SMOTE white paper: https://www.jair.org/media/953/live-953-2037-jair.pdf
A set of simple Python scripts for pre-processing large files
Code for Youtube video: https://www.youtube.com/watch?v=dTRDZBltCTg&list=UUq4pm1i_VZqxKVVOz5qRBIA&index=1
Easy to use Geocoding tool for Excel.
Ensemble
John Langford's original release of Vowpal Wabbit -- a fast online learning algorithm
Kaggle Competion - Multi-label prediction
Peer Assessment 1 for Reproducible Research
Getting and Cleaning Data - John Hopkins
eXtreme Gradient Boosting (Tree) Library
Plotting Assignment 1 for Exploratory Data Analysis
Repository for Programming Assignment 2 for R Programming on Coursera
Function wrapper for Caret's dummyVars to quickly break out every factor for a given data.frame column. Handles one column formula on 2 or more levels and returns altered data.frame with new columns minus original.
The Leek group guide to data sharing
repository for JH/Coursera class 'The Data Scientist’s Toolbox'
Public repository.
Kaggle competition
(C#) An interesting enhancement is adding the bid and ask size differential horizontally next to each candlestick. The idea is easy to grasp, when excessive offers overcome bids, the price tends to fall, and, in the opposite scenario, rise. This is obvious on the quote monitor but tedious to follow. Displaying it on the chart is clear and may enhance existing trading systems.
(C#) For the semi-discretionary trader, there is a feature lacking in NinjaTrader which is easily remedied with a little C# code: customizable buttons to trigger strategies. If you watch the charts to initiate a trading decision but use complex entry, positioning and stop methodology, being able to trigger a strategy manually may be of great use. For example, you can create two buttons tied to a dynamic strategy that buys a tick above the high or sells a tick below the low of the last closed bar. Before sending out the order, you can calculate the size dynamically according to the market, buying power, and/or size of the signal bar. Caution! This code will send orders in NT, do not use if you don't know what you are doing!! This is as is, the risk is yours.
(R) Calculates distance in miles between two US zip codes using the R language
(Python) They know who you're going to vote for just by looking at your face… (facial recognition using Python)
(Python) Logistic regression to predict the S&P 500 using Python
(Python) Using Random Patterns to Predict the S&P 500