Practice_4

---
output:
  pdf_document
---

# This is our first autogenerated report

> Here we collect data from Kimono Labs API and produce autogenerated minireport.

Here you can see activity on publishing news on [RIA Lenta](http://ria.ru/lenta).

```{r, include=FALSE}
api.name <- '1xl232p8'
api.key <- 'DTndIua4tfpz5kxITOILI6FvmGV8xw9g'
Sys.setlocale(category = "LC_ALL", locale = "C")
library(RCurl)
library(rjson)
library(ggplot2)
library(RColorBrewer)
```

```{r, echo=FALSE}
json.url <- paste0('https://www.kimonolabs.com/api/', api.name, '?apikey=', api.key)
json.file <- getURL(json.url, .encoding = 'UTF-8',
                    .opts = list(ssl.verifypeer = FALSE))
json.data <- fromJSON(json.file)
df <- data.frame()
for(i in 1:json.data$version){
  csv.url <- paste0('https://www.kimonolabs.com/api/csv/', i, '/', api.name, '?apikey=', api.key)
  csv.file <- getURL(csv.url, .encoding = 'UTF-8', .opts = list(ssl.verifypeer = FALSE))
  csv.rows <- read.csv(textConnection(csv.file), skip=1)
  df <- rbind(df, csv.rows)
}
df <- unique(df)
df$timestamp <- strptime(as.character(df$timestamp), format = '%H:%M %d.%m.%Y')
df$hour <- df$timestamp$hour
df$day <- strftime(df$timestamp, format='%d %b')
df <- aggregate(df, by = list(df$day, df$hour), FUN = length)
df <- df[, 1:3]
names(df) <- c('Date', 'Hour', 'Amount')
specPal <- colorRampPalette(rev(brewer.pal(11, "Spectral")))
plot <- ggplot(df, aes(x = Hour, y = Date, fill = Amount))
plot <- plot + geom_tile()
plot <- plot + scale_fill_gradientn(guide="colourbar", colours = specPal(100), guide_legend(title="Activity"))
plot <- plot + scale_x_discrete("Hour", limits=0:22, labels=0:23)
plot <- plot + scale_y_discrete("Day")
plot <- plot + coord_equal()
plot <- plot + theme_bw()
plot <- plot + theme(legend.position="bottom", legend.background = element_rect(colour = "grey"))
plot <- plot + ggtitle(expression("Publication frequency"))
print(plot)
```

## Other abilities

### Use lists:

* Item 1
* Item 2
* etc.

### Or multiple headers:

#### Level-4 header

### Math is also possible:

$EngagementRate = 100 \times \frac{likes + shares + comments}{Followers} (\%)$

## Conclusion

Use R and have fun.