swayson

##VGG19 model for Keras

This is the Keras model of the 19-layer network used by the VGG team in the ILSVRC-2014 competition.

It has been obtained by directly converting the Caffe model provived by the authors.

Details about the network architecture can be found in the following arXiv paper:

Very Deep Convolutional Networks for Large-Scale Image Recognition

K. Simonyan, A. Zisserman

swayson

import pandas as pd
from github import Github

g = Github("username", "password")

final = ({'url':r.html_url , 'name': r.name} for r in g.get_user().get_starred())
pd.DataFrame(final).to_excel('Github Stars 20160101.xlsx')

swayson

starts_with <- function(vars, match, ignore.case = TRUE) {
  if (ignore.case) match <- tolower(match)
  n <- nchar(match)

  if (ignore.case) vars <- tolower(vars)
  substr(vars, 1, n) == match
}

ends_with <- function(vars, match, ignore.case = TRUE) {
  if (ignore.case) match <- tolower(match)

swayson

mylist <- list(structure(list(Hit = "True", Project = "Blue", Year = "2011", 
Rating = "4", Launch = "26 Jan 2012", ID = "19", Dept = "1, 2, 4"), .Names = c("Hit", 
"Project", "Year", "Rating", "Launch", "ID", "Dept")), structure(list(
Hit = "False", Error = "Record not found"), .Names = c("Hit", 
"Error")), structure(list(Hit = "True", Project = "Green", Year = "2004", 
Rating = "8", Launch = "29 Feb 2004", ID = "183", Dept = "6, 8"), .Names = c("Hit", 
"Project", "Year", "Rating", "Launch", "ID", "Dept")))

dfs <- lapply(mylist, data.frame, stringsAsFactors = FALSE)
library(dplyr)

swayson

minmax_scaler <- function(x, a, b) {
    "
    x: data. numeric vector of values to be scaled
    a: desired minimum after scaling takes place
    b: desired maximum after scaling takes place
    
    e.g. f(c(1,2,3,4), 1, 17)
    [1]  1.000000  6.333333 11.666667 17.000000
    "
    (((b - a)*(x - min(x))) / (max(x) - min(x))) + a

swayson

def select(dataframe, columns, keep_others=True):
    ''' Re-order or select columns. If keep_others, then it is simply re-ordered else it will select columns'''
    cols = set(dataframe.columns)
    if keep_others:
        others = list(cols.difference(columns))
        reordered = columns + others
        return dataframe[reordered]
    else:
        return dataframe[columns]

swayson

def search_item(dataframe, name, query, na=False, case=False, regex=True):

    idx = pd.Series([False]*len(dataframe))

    # For each item in the query look for the item and collect the documents ids it pertains to
    for q in query:
        matches = dataframe[text_column].str.contains(q, na=False, case=False, regex=True)

swayson

swayson

1. Ditch Excel (for real)

    in2csv file1.xls > file1.csv

2. Conquer fixed-width formats

    in2csv -f fixed -s schema.csv data.fixed > data.csv

3. Find cells matching a regular expression

csvgrep -c phone_number -r "\d{3}-123-\d{4}" data.csv > matching.csv

swayson

---
title: 'Going deeper with dplyr: New features in 0.3 and 0.4'
output: html_document
---

## Introduction

In August 2014, I created a [40-minute video tutorial](https://www.youtube.com/watch?v=jWjqLW-u3hc) introducing the key functionality of the dplyr package in R, using dplyr version 0.2. Since then, there have been two significant updates to dplyr (0.3 and 0.4), introducing a ton of new features.

This document (created in March 2015) covers the most useful new features in 0.3 and 0.4, as well as other functionality that I didn't cover last time (though it is not necessarily new). My [new video tutorial](https://www.youtube.com/watch?v=2mh1PqfsXVI) walks through the code below in detail.