Analytics Sandbox

Overview and Introduction Driven by the desire to do something as fun and interesting as this post by the folks at Decision Science, I decided to dust off some previous work done on Louisville Metro Animal Services Animal [“LMAS”] Intake and Outcome data (you can get it here). The problems are sort of the same at the core - while the linked work focuses on customer churn, this analysis will focus on what information tends to predict a given animal will be euthanized at LMAS using Artificial Neural Networks (“ANN”).

Background The purpose of this tutorial is to walk through using d3.js within an Rmd document that is then rendered as an HTML page with a Hugo static site generator. That was probably the most technical sentence of this entire write-up, so let’s all breath a sigh of relief now that we’re through with it. This is a quick release, so I’m not going to go into much detail about the process to get the data and what it fully represents just yet - I’m saving that for a deeper post with a “tbd” release date at this point.

Tutorial’s purpose Manipulating data and building visuals is the bread and butter of the R language - but what can get lost in the fold is the ability to easily and quickly push analysis out for folks to see. Using knitr and rmarkdown makes that significantly easier, and exists at the press of a button. In this tutorial, we’re going to press those buttons. What are the packages? There are a ton of ways to get information from R out into the world - ranging from static Word or PDF documents all the way to interactive web applications.

Today’s purpose To walk through the tidyverse and become familiar with the functions that it provides the user and how it can make your life so much better to use. What the heck even is a tidyverse? tidyverse is a collection of R packages providing an all-inclusive resource for data science (well, almost). When you library(tidyverse) the following packages are loaded up as dependencies: Package Description ggplot2 Graphics-building package based on The Grammar of Graphics mapping aesthetics to data visually tibble Wrapper on traditional dataframes allowing for better printing and viewing data tidyr Package to easily and quickly get your data into a rectangular format (if not already in one) readr Super fast package for reading in rectangular data (like csv, tsv, and fwf) purrr A set of tools for working with functions and vectors for functional programming techniques dplyr Quickly and efficiently manipulate data with a Grammar of Manipulation (just made that up) Within each of these there are a handful of other dependencies - not all of which I am going to talk through.

Background This is a working document to help understand the first stage of this project - the data grab. It is also a helpful learning tool for myself, as this is the first real spider that I’ve created to pull data from a website. Setup The initial setup is quick, simply cd into the directory that you want to work in and run - scrapy startproject whiskey_reviews Easy enough.