Create interactive bar chart with shared data filtered by time range

user_1977053 0
Viewed 1456

I want to create an interactive bar chart that lets users filter observations based on a range of values, and then renders counts per class for the selected time period dynamically. Since the filtered data needs to be available for numerous such graphs, I thought a combination of crosstalk and plotly/ggplot might prove valuable.

I attached a reprex further below that uses shared data and filtering functionality from crosstalk to allow for the dynamic filtering part. When I knit the document, the bar chart renders nicely as long as the full range of values is selected (default).

full

However, the plotting region becomes empty for any other, ie. user-adjusted range.

partial

What exactly am I missing here? I assume there must be a difference between full and filtered shared datasets that ggplotly() cannot handle proberly. Is there maybe another approach that I could follow to achieve my goal?

Here's the content of my .Rmd file:

---
title: mpg class counts filtered by time period
output: html_document
---

```{r echo = FALSE, message = FALSE, warning = FALSE}
library(crosstalk)
library(plotly)

# Wrap data frame in SharedData
sd = SharedData$new(mpg)

# Create a filter input
filter_slider("Year", "Year", sd, column = ~ year, step = 1, width = 250)

# Render graph
bscols(
  ggplotly(
    ggplot(aes(x = class), data = sd) + 
      geom_bar()
  )
)

```
crosstalk ggplotly r-plotly r visualization
2 Answers

I think this may be because "Crosstalk currently only works for linked brushing and filtering of views that show individual data points, not aggregate or summary views (where “observations” is defined as a single row in a data frame). For example, histograms are not supported since each bar represents multiple data points; but scatter plot points each represent a single data point, so they are supported." official doc

If you change it to point plot, it seems to be working.

---
title: mpg class counts filtered by time period
output: html_document
---

```{r echo = FALSE, message = FALSE, warning = FALSE}
library(data.table)
library(crosstalk)
library(plotly)

# Wrap data frame in SharedData
sd = SharedData$new(mpg)

# Create a filter input
filter_slider("Year", "Year", sd, column = ~ year, step = 1, width = 250)

# Render graph
bscols(
  ggplotly(
    ggplot(aes(hwy, cty), data = sd) + 
      geom_point()
  )
)

```
user_123477080

Would following work for you? If you want to filter for dates, you might want to have a look at plotly::rangeslider.

library(tidyverse)
library(plotly)

df <- crosstalk::SharedData$new(mpg)$data() %>%
  group_by(year, class) %>%
  count() %>%
  mutate(year = as.factor(year))
  
df %>%
  plot_ly(x = ~class, y = ~n, color = ~year) %>%
  add_bars() %>%
  layout(barmode = "stack")
user_123477080
Related