binning-in-tidyverse.Rmd

binning-tidyverse.Rmd

---
title: "binning-in-tidyverse"
output: rmarkdown::github_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r load-packages, echo=FALSE}
library(tidyverse)
```

I'll just generate 1000 rows of fake data to use as an example -- use your own dataframe instead of `data` here! 

```{r make-data}
id <- paste0('person', 1:1000, sep="")
time <- rexp(1000, rate=1)
data <- data.frame(id, time)
```
Here's what the first 20 rows look like:

```{r show-data}
head(data, 20)
```

## Binning

Let's use `mutate` to create a new column, `binned_time`. This will show the bin that each response time falls in.
If the width of each bin is equal (say, 0.1 of a sec each), you can use `ceilf` or `floor` for this:

```{r equal-bins}
data <- data %>% mutate(binned_time = floor(time*10)/10)
head(data, 20)
```
(Let me know if you want unequal bin sizes -- this is still possible just a tiny bit more complicated!)

## Counts

Let's count how many observations fall into each bin.

```{r tally}
data %>% count(binned_time)
# NB: This is the same as
# data %>% group_by(binned_time) %>% tally()
```

## Splitting data by bins

Maybe you do actually want to split the data up by the binned time!
For instance, maybe you want to save each set of binned times separately so that you can
run different analyses on each of them later on. 
In this case, you can use `group_split()`. The output is a list.

```{r split-data}
bin_list <- data %>% group_by(binned_time) %>% group_split()
bin_list[[3]]
```