mjamesruggiero

;; assumes that you have a JSON file 
;; corresponding to /api/v1/user/exercises?username=<your-username>
(ns khan.core
  (:require [clojure.data.csv :as csv]
            [clojure.java.io :as io]
            [clojure.data.json :as json]))

(defn get-json
  [filepath]
  (with-open [in-file (io/reader filepath)]

mjamesruggiero

#! /usr/bin/env racket
#lang racket

;;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
;;         pivotal
;;
;; what it does: pulls the current iteration,
;; echoes short story summary
;; binned by the first owner's name
;;

mjamesruggiero

;; note that project loads this in :dependencies
;; [org.xerial/sqlite-jdbc "3.7.2"]]

(ns piney.pagerduty
  (:require [piney.utils :refer :all]
            [clojure.java.jdbc :refer :all]
            [clojure.java.io :as io]
            [clojure.edn :as edn]))

(defn load-config

mjamesruggiero

1:1 questions

Always end with an assignment

What can I hold you accountable for next time we talk?

What can I be accountable to you for the next time we talk?

Short term

How is [project] going? What could we do to make it better?

Is there anything blocking you from getting your work done?

Are there any projects you’d really like to work on if you were given the opportunity?

mjamesruggiero

package com.mjamesruggiero

import scalaz.State

object StateMonad {

  type StateCache[+A] = State[Cache, A]
  trait ConfigService {
    def value(k: String): StateCache[ConfigValue]
  }

mjamesruggiero

Personalization and Recommenders with Content-Based Approaches

Jake Mannix, Applied Machine Learning Engineer, Twitter

• Does collaborative filtering == RecSys?
• cautions that recommenders are not just user->products
• you might actually recommend groups to a user, user to a groups, candidates to a posting (example: Linkedin)

math techniques:

• singular value decomposition or SVD

mjamesruggiero

Joseph Gonzalez, Co-Founder at GraphLab

Graphs encode the relationship between entities and are essential to data mining and machine learning.

Example: predicting user behavior

• as a base technique that does a classification re: political bias
• you can estimate the political bias and take that idea and apply it to a large user base
• the vast majority of the users may not post frequently, but you have the follower structure, the graph, so you can use a conditional random field

"triangle counting"

mjamesruggiero

Pandora talk re:recommendation systems

Eric Bieschke, Chief Scientist @pandora

• Expresses great love for A/B testing
• "the big advantage about having a lot of data is that you can do experiments with real data, real users"

The importance of metrics

• how you judge experiments shapes where you are headed
• choose the wrong measuring stick and you wind up in the wrong place

mjamesruggiero

large scale deep learning

Quoc V.Le - Google

parallel neural networks at Google scale

• machine learning requires domain knowledge from human experts
• we want to move beyond hiring domain experts; it would be good to have machines create features rather than human experts

deep learning:

mjamesruggiero

# Fri Mar 15 21:31:59 PDT 2013 
# want a histogram of queries
import re
import logging
import csv

def clean(line):
  return remove_newlines(remove_times(sub_digits(line)))

def sub_digits(old_string, subst="x"):