Rob Coleman RobColeman

I build smart applications and make applications smarter with machine learning, statistics, and appropriate algorithms.

RobColeman / pdc_dtf.py

Last active August 29, 2015 14:06 — forked from agramfort/pdc_dtf.py

	#!/usr/bin/env python
	# -- coding: utf-8 --

	"""Implements Partial Directed Coherence and Direct Transfer Function
	using MVAR processes.

	Reference
	---------
	Luiz A. Baccala and Koichi Sameshima. Partial directed coherence:
	a new concept in neural structure determination.

RobColeman / d3_course_dataset_meteorites

Last active August 29, 2015 14:18

d3_course_dataset_meteorites

	I have chosen the meteorites dataset from The Meteorite Society, hosted at http://visualizing.org/datasets/meteorite-landings .

	It has a couple interesting things going for it.

	- It's only about 3.1MB, not very big, but not to small in size.
	- It's got a lot of different data-types, with intersting relationships to play with (variable-types)
	- Ratio (Mass, year)
	- Qualitative - Interval (lat, long)
	- Catagorical non-ordinal (class/type, namtype)
	- It's already clean

RobColeman / index.html

Created April 3, 2015 20:33

meteorites_module2_exercize

	<!DOCTYPE html>
	<html lang="en">
	<head>
	<meta charset="utf-8">
	<title>Loading CSV Data with D3</title>
	<script type="text/javascript" src="http://d3js.org/d3.v3.js"></script>
	</head>
	<body>

	<p>Not much to see here; try looking in the console!</p>

RobColeman / gist:2babd0fad3d7f01f3d7e

Last active August 29, 2015 14:21

	for x in "asdfasdfas2343asdf232dfasdf": print(x.isdigit())


	def how_many_numbers_in_the_word(word):
	counter = 0
	for x in word:
	if x.isdigit():
	counter = counter + 1
	return counter

RobColeman / count_nucleotides.py

Last active October 27, 2015 18:39

Counting nucleotides from a sequence

	def count_nucleotides_by_type(sequence):
	# use a dictionary to keep track of the counts of each nucleotide type in the sequence
	counts = {
	"A": 0,
	"C": 0,
	"G": 0,
	"T": 0,
	"length": 0
	}

RobColeman / some_algorithms.py

Created November 9, 2015 22:19

Some Algorithms In Python

	# binary search
	def binary_search(x, arr):
	"""
	x is a integer,
	arr is an array of integers, in sorted order

	Binary search will tell you if x is contained in arr

	A binary search bisects the array, recursively, to search smaller and smaller
	sub-sections of the array until it either, finds the value or finds no more candidate

RobColeman / TDigest.scala

Last active January 18, 2018 17:21

A fast, serializable, Scala implementation of tDigest. Thrown together for a Spark project.

	import java.nio.ByteBuffer

	import org.apache.spark.{SparkContext, SparkConf}
	import org.apache.commons.math3.distribution.ExponentialDistribution
	import scala.collection._
	import scala.collection.generic.CanBuildFrom
	import scala.util.Random

	case class Centroid(var mean: Double, var count: Long) extends Ordered[Centroid] with Serializable {

RobColeman / TreeDigestHelper.scala

Last active March 26, 2019 16:09

Helpers for TDunnings Java TDigest library

	package com.preact.platform.math.models

	import java.lang.System._
	import java.nio.ByteBuffer

	import org.apache.spark.rdd.RDD
	import org.apache.spark.{SparkContext, SparkConf}
	import org.apache.commons.math3.distribution.ExponentialDistribution
	import org.apache.commons.math3.distribution.NormalDistribution
	import com.tdunning.math.stats.TreeDigest

RobColeman / ApproximateDistribution.scala

Last active January 29, 2016 01:06

An Approximate Distribution wrapper class for TDunnings java TDigest library

	package com.preact.platform.math.models

	import java.lang.System._
	import java.util
	import com.tdunning.math.stats.{Centroid, TreeDigest}
	import org.apache.commons.math3.distribution.NormalDistribution
	import org.apache.spark.rdd.RDD
	import org.apache.spark.{SparkContext, SparkConf}

	import scala.collection.immutable.IndexedSeq

RobColeman / NewFeatureComputers.scala

Created August 24, 2016 05:35

	package com.chartboost.adrel.preprocessing.featureComputers

	import com.chartboost.adrel.dataModels.EcpmPredictionRequest
	import com.chartboost.adrel.util.JsonSaving
	import org.apache.spark.mllib.linalg.{Vector => SparkVector, Vectors => SparkVectors}
	import scala.util.hashing.MurmurHash3

	/**
	* The meta-data for computing features. In this format to easily save and load with models.
	*