cpi cut number

cpi-cut-number

```{r include = FALSE}
library(tidyverse)
library(tidyquant)
library(timetk)
library(readxl)
library(plotly)
library(scales)
library(formattable)
library(fredr)
library(broom)

```
 
 Suppose we didn't wish to choose levels, but rather wished to have our levels chosen so that an equal number of observations falls in each bucket. To accomplish this, we call the `cut_number()` function from `ggplot2`. We tell `cut_number()` function how many buckets to create, and the function creates them such that each has an equal number of observations. 

```{r}
cpi <- 
  "CPIAUCSL" %>% 
tq_get(get = "economic.data", from = "1979-01-01") %>% 
  select(date, cpi = price) %>% 
  mutate(cpi_month_change = cpi/lag(cpi, 1) - 1,
         cpi_year_change = cpi/lag(cpi, 12) - 1,
         cpi_5_year_change = cpi/lag(cpi, 60) - 1)

cpi_monthly_yoy_change <- 
cpi %>% 
  select(date, cpi_year_change) %>% 
  mutate(quarter_year= str_glue(" {year(date)} Q{quarter(date)} ") %>% 
           as.yearqtr())
```


```{r}
  cpi_monthly_yoy_change %>% 
  drop_na() %>% 
  mutate(cpi_yoy_bucket = cut_number(cpi_year_change, n = 7)) %>% 
  group_by(cpi_yoy_bucket) %>% 
  count()
```

Notice how each bucket has 71 or 72 observations, and the ranges are quite precise. Since we eventually want to chart and use these levels, we want better labels than what's displayed currently. We created a custom function called `cut_label_fun()` to accomplish this. 

We use a custom function called cut_label_fun(). 

And here are the new results: 

```{r}
cpi_monthly_yoy_change %>% 
  drop_na() %>% 
  mutate(cpi_yoy_bucket = cut_number(cpi_year_change, n = 7)) %>% 
    cut_label_fun(bucket_col = cpi_yoy_bucket) %>% 
  group_by(cpi_yoy_bucket, label) %>% 
  count()
```

No change in substance, but we have a column called `label` with more aesthetically pleasing descriptions for our buckets. Now let's plot a visualization of those buckets and how actual CPI changes fall within them.

```{r}
# inflation_bucketed_equal <- 
  cpi_monthly_yoy_change %>% 
  mutate(cpi_yoy_bucket = cut_number(cpi_year_change, n = 7)
  ) %>%
  filter(!is.na(cpi_yoy_bucket),
         date >= "1980-01-01") %>% 
    cut_label_fun(bucket_col = cpi_yoy_bucket) %>% 
  ggplot(aes(x = label, y = cpi_year_change, color = label)) +
    geom_jitter()  +
  scale_y_continuous(labels = percent, breaks = pretty_breaks(n = 10)) + 
  theme_minimal() +
  labs(title =  "Distribution of Monthly CPI YoY", 
       subtitle = "Equal Number obs per bucket",
         color = 
         "Inflation Bucket", x = "", y = "Actual YoY Change CPI",
       caption = "Source: FRED Data and author calcs") +
  theme(legend.position = "right",
        plot.title = element_text(hjust = .5),
       plot.subtitle = element_text(hjust = .5) ) +
    scale_x_discrete(guide = guide_axis(n.dodge = 2)) 
```

Hi. Thanks for creating this really interesting chart. Question: I am not seeing your definition of the custom function "cut_label_fun". Getting this error message:
Error in cut_label_fun(., bucket_col = cpi_yoy_bucket) :
could not find function "cut_label_fun"