---
title: "BART and json data"
output:
  word_document: default
  html_notebook: default
  pdf_document: default
---

Today we are going to take a look at the BART website and look at the visualizations presented on the website [analytics.bart.gov](https://analytics.bart.gov/bart).

We will take a look at *json* formated data, which is a very common data format for sharing data on the internet.

## Directions:

1. Create an R Project call BART.
2. Download the [devices](https://www.bart.gov/sites/default/files/analytics/devices.json) data.
3. Try the following R code to make the bar graph for the devices used to access the BART website.

Load the libraries used.

```{r}
library(tidyverse)
library(jsonlite)
```

Read in the json data.

```{r}
json_file <- "devices.json"

json_data <- fromJSON(json_file)
json_data
```

Fix the visits column so that it is a numeric value.

```{r}
json_data <- json_data %>% mutate(visits = as.numeric(visits))
json_data
```

Make a table of the most recent day's data.

```{r}
json_data %>% filter( date == "20190922" ) %>%
  group_by(device) 
```

Make a bar plot.

```{r}
json_data %>% filter( date == "20190922" ) %>%
  ggplot(aes(y = visits, x = device)) +
  geom_bar(stat="identity")
```

Make a bar plot using percentages.

```{r}
json_data <- json_data %>% filter( date == "20190922" ) %>%
  mutate(visits_percentage = 100*visits/sum(visits))
json_data
```

```{r}
json_data %>% ggplot(aes(y = visits_percentage, x = device)) +
  geom_bar(stat="identity") +
  coord_flip()
```