jmcastagnetto

# bug in sjt.xtab()

```{r}
require(sjPlot)
set.seed(12345)
var1 <- as.factor(sample(c("ok", "not ok"), 1000, replace=TRUE, prob=c(0.2, 0.8)))
var2 <- as.factor(sample(c("foo", "bar", "boing", "kaboom", "plonk"), 1000, replace=TRUE))
var3 <- sample(2:20, 1000, replace=TRUE)
xtab1 <- sjt.xtab(var2, var3, var1, variableLabels=c("var2", "var3", "var1"),
                  showSummary=FALSE, showLegend=FALSE, no.output=TRUE)

jmcastagnetto

---
title: "Analysis of the data on Presidents of Peru"
author: Jesus M. Castagnetto
output:
  html_document:
    toc: true
    theme: readable
    highlight: tango
---

jmcastagnetto

# This is quick and dirty code that tries to reproduce the analysis and graph from the post:
# "Per capita consumption of mozzarella cheese (US) correlates with 
# Civil engineering doctorates awarded (US)"
# URL: http://www.tylervigen.com/view_correlation?id=3890

# get the data from the article
library(XML)
url <- "http://www.tylervigen.com/view_correlation?id=3890"
tables <- readHTMLTable(url)
# we need table 3, minus a couple of extra columns

jmcastagnetto

Script to be used with Pingdom's custom HTTP check.

Inspired by the code described in http://jonsview.com/how-i-use-pingdoms-http-custom-feature

Warning:

As of 20114-08-07, if you are using gzip compression in your server, creating a check pointing to an URL with the output of this script will fail, as the code been used by Pingdom does not yet grok compression.

The solution (at least for Apache), is to explicitly exclude from compression the path to the script in your server config, e.g.:

jmcastagnetto

---
title: "The human health and financial impact of storms (1950-2011)"
author: "Jesus M. Castagnetto"
date: "2014-08-22"
output: html_document
---

**Course: Reproducible Research (Coursera), Assignment 2**  

```{r}

jmcastagnetto

---
title: "Part 1: Simulation using the exponential distribution"
author: "Jesus M. Castagnetto"
date: "2014-08-23""
output:
    pdf_document:
        highlight: zenburn
fontsize: 10pt
geometry: margin=0.8in
---

jmcastagnetto

# Kahan summation
# ref: http://rosettacode.org/wiki/Kahan_summation

# Try using the task rules
# 1. Do all arithmetic in decimal to a precision of six digits.
options(digits=6)

# 2. Write a function/method/procedure/subroutine/... to perform Kahan summation
# on an ordered collection of numbers, (such as a list of numbers).
kahansum <- function(x) {

jmcastagnetto

<?php
// ref: http://rosettacode.org/wiki/Kahan_summation

// 1. Do all arithmetic in decimal to a precision of six digits.
// PHP does not have a way of restricting overall precision

// 2.  Write a function/method/procedure/subroutine/... to perform Kahan
// summation on an ordered collection of numbers, (such as a list of numbers). )
function kahansum($input) {
    $sum = $c = 0;

jmcastagnetto

require(RCurl)
require(XML)

# Let's make a special class
NISTBeaconResponse <- function (ts) {
    if(!is.integer(ts) &
       !inherits(ts, "POSIXct") &
       !inherits(ts, "POSIXlt")) {
        stop("We expected a unix timestamp as an integer or a POSIXct or POSIXlt value")
    }

jmcastagnetto

> library(dplyr)

> ft <- mtcars %>% group_by(gear) %>% summarise(freq=n())

> ft
Source: local data frame [3 x 2]

  gear freq
1    3   15
2    4   12