---
title: "Overview of sparklyr.flint"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{sparklyr-flint}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

[sparklyr.flint](https://github.com/r-spark/sparklyr.flint) is a sparklyr extension that aims to make [Flint](https://github.com/twosigma/flint) time series functionalities easily accessible from R.

Analyzing time series data using sparklyr.flint usually entails the following:

 * Importing time series data into a Spark dataframe if it is not in Spark already
 * Conversion of Spark dataframe into a TimeSeriesRDD, a time-series-aware data structure
 * Analyzing the resulting TimeSeriesRDD using Flint time series functionalities -- at the moment, sparklyr.flint supports most of the summarizers implemented by [Flint](https://github.com/twosigma/flint) and supports summarization of both time windows and cycles