description_text_analysis.Rmd

---
title: "Description Text Analysis"
output: html_notebook
---

_Reminder: Don't forget to set your working directory._

This script reads in the campaign spending dataset from 2016 and counts the 
ngrams found in the spending description vector. The counts are stored in the 
variable `words`. The ngram counts can be used to narrow down spending 
transactions into discreet categories.

```{r}
# Load packages
library(tidyverse)
library(tidytext)
```

```{r}
# Read data from data.world
expend_16 <- read.csv("https://query.data.world/s/4hhflof4h6w5zsnrdgeux239z")
```

```{r}
# Convert into tidy dataset with a words from categories
data("stop_words")

words <- expend_16 %>%
  unnest_tokens(word, cat, token = "ngrams", n = 2) %>%
  anti_join(stop_words, by = "word") %>%
  count(word, sort = TRUE)
```