TidyBlock #作る

🧱 TidyBlock

A Visual, Node-Based GUI for dplyr Data Wrangling

Build tidyverse pipelines by dragging and clicking — no coding required.





📖 Overview

TidyBlock is an R Shiny application that lets you visually construct dplyr data-wrangling pipelines through an intuitive, block-based interface. Instead of writing R code manually, you add pipeline modules (Filter, Select, Mutate, etc.) as visual cards and configure them through dropdowns and inputs. TidyBlock generates the equivalent R code in real time.

✨ Key Features
FeatureDescription🔍 Advanced Filter Build complex filter conditions with AND/OR groups and nested logic
📋 Select Choose which columns to keep
➕ Mutate Create new columns with custom R expressions
📊 Group By Group data by one or more variables
📈 Summarize Aggregate data with mean, sum, median, sd, min, max, n, n_distinct
↕️ Arrange Sort data ascending or descending
🔗 Join Left/Inner/Right/Full join with auto-detected keys
📉 Plot Create ggplot2 visualizations (Point, Bar, Line, Boxplot) with themes
💻 Live Code Preview See the generated dplyr pipeline code update in real time
📊 Live Data Preview Instantly see filtered/transformed data as you build
↔️ Resizable Cards Resize module cards with drag handles


🖼️ Screenshots


After launching the app, load the demo Iris dataset and add pipeline modules from the sidebar.

Pipeline with Filter, Select, and Mutate
┌─────────────────────────────────────────────────────┐ │ TidyBlock │ ├──────────┬──────────────────┬────────────────────────┤ │ Toolbox │ Data Preview │ Filter (Advanced) │ │ │ │ [AND ▼] [+ Rule] │ │ ▶ Filter │ Sepal.Length... │ Sepal.Length == 5.1 │ │ ▶ Select │ ├────────────────────────┤ │ ▶ Mutate │ │ Select │ │ ▶ GroupBy│ │ [Sepal.Length, ...] │ │ ▶ Summ. │ R Code Preview ├────────────────────────┤ │ ▶ Arrange│ iris |> │ Mutate │ │ ▶ Join │ filter(...) |> │ new_var = log(...) │ │ │ select(...) |> │ │ │ ▶ Plot │ mutate(...) │ │ └──────────┴──────────────────┴────────────────────────┘



🚀 Getting Started

PrerequisitesR >= 4.3.0
Conda (Miniconda or Anaconda) — recommended for reproducible environments

Installation

Option 1: Using Conda (Recommended)
# 1. Clone the repository git clone https://github.com/amufaamo/tidyblock.git cd tidyblock # 2. Create the Conda environment conda env create -f environment.yml # 3. Activate the environment conda activate void # 4. Launch the app R -e "shiny::runApp()"



Option 2: Manual R Package Installation
# 1. Clone the repository git clone https://github.com/amufaamo/tidyblock.git cd tidyblock # 2. Install dependencies in R R -e "install.packages(c('shiny', 'bslib', 'tidyverse', 'rlang', 'DT', 'shinyAce', 'shinyjqui', 'bsicons', 'glue', 'pkgload'))" # 3. Launch the app R -e "shiny::runApp()"



The app will open at http://127.0.0.1:xxxx (port is shown in the console).

Quick Start GuideLoad Data — Click "Load Demo Data (Iris)" in the sidebar, or upload your own CSV file.
Add Modules — Click any button under "Pipeline Tools" (Filter, Select, Mutate, etc.) to add a module card.
Configure — Set options in each module card (e.g., select columns, set filter conditions).
View Results — The "Data Preview / Result" table and "R Code Preview" update in real time.
Copy Code — Copy the generated R code from the preview panel to use in your own scripts.

📦 Project Structure
tidyblock/ ├── app.R # Entry point — launches the Shiny app ├── DESCRIPTION # R package metadata (version, dependencies) ├── NAMESPACE # R namespace declarations ├── CHANGELOG.md # Version history ├── environment.yml # Conda environment definition ├── iris.csv # Demo dataset ├── R/ │ ├── app_ui.R # Main UI layout (sidebar + main panel) │ ├── app_server.R # Main server logic (pipeline execution engine) │ ├── mod_import.R # Data import module (CSV upload, demo data, preview) │ ├── mod_filter.R # Advanced filter module (AND/OR groups, rules) │ ├── mod_select.R # Column selection module │ ├── mod_mutate.R # Column creation module │ ├── mod_group_by.R # Group-by module │ ├── mod_summarize.R # Summarize/aggregate module │ ├── mod_arrange.R # Sort/arrange module │ ├── mod_join.R # Join module (left, inner, right, full) │ ├── mod_plot.R # ggplot2 visualization module │ └── utils_codegen.R # Pipeline code generation utilities ├── man/ # R documentation (auto-generated) ├── tests/ # Unit tests (testthat) └── DEBUGGING.md # Debugging guide for developers



🔧 Module Reference

Filter (Advanced)

Build complex filter conditions with nested AND/OR logic.
# Example generated code: iris |> filter((Sepal.Length > 5 & Species == 'setosa') | Petal.Width < 0.3)

AND/OR Groups: Combine conditions with logical operators
Nested Groups: Create sub-groups for complex logic (e.g., A AND (B OR C))
Operators: ==, !=, >, <, >=, <=, %in%

Select

Choose columns to keep in the dataset.
iris |> select(`Sepal.Length`, `Sepal.Width`, `Species`)



Mutate

Create new columns with R expressions.
iris |> mutate(`sepal_ratio` = `Sepal.Length` / `Sepal.Width`)

Helper Functions: Quick insert for log(), sqrt(), abs(), round(), paste0(), as.numeric(), as.character(), ifelse()

Group By + Summarize

Group data and compute aggregations.
iris |> group_by(`Species`) |> summarize(`mean_sl` = mean(`Sepal.Length`), .groups = 'drop')

Functions: mean, sum, median, sd, min, max, n, n_distinct
Optional Secondary Statistic: Add a second aggregation (e.g., mean ± sd)

Arrange

Sort data by a column.
iris |> arrange(desc(`Sepal.Length`))



Join

Combine datasets with dplyr join functions.
iris |> left_join(other_data, by = c("Species"))

Join Types: Left, Inner, Right, Full
Auto-detected Keys: Common columns between datasets are automatically suggested

Plot

Create ggplot2 visualizations.
iris |> ggplot(aes(x = `Sepal.Length`, y = `Sepal.Width`, color = `Species`)) + geom_point() + theme_minimal()

Geometries: Point, Bar (auto-switches to geom_col), Line, Boxplot
Optional Error Bars: geom_errorbar() with ymin/ymax
Themes: gray, bw, linedraw, light, dark, minimal, classic, void

🏷️ Version History

v1.0.0 (2026-02-15) — Initial Stable Release

The first public release with all core data-wrangling modules fully functional.

What's included:✅ All 8 pipeline modules (Filter, Select, Mutate, Group By, Summarize, Arrange, Join, Plot)
✅ Advanced Filter with AND/OR nested group logic
✅ Real-time code generation and data preview
✅ Resizable module cards (shinyjqui)
✅ Reactive pipeline execution with error handling
✅ Server-side DataTable rendering for performance
✅ Conda environment for reproducible setup

Bug Fixes from Development:Fixed reactive circular dependency that caused UI to freeze after loading data
Fixed DT::renderDT(server = FALSE) blocking the Shiny event loop
Re-enabled jqui_resizable() on all module cards
Restored full Filter (Advanced) UI from test placeholder

v0.1.0 (2026-02-04) — Development Preview

Internal development version. Not publicly released.

What was included:⚠️ Basic project scaffolding
⚠️ Import module with CSV upload and demo data
⚠️ Placeholder module files (not fully functional)
❌ Pipeline tools did not work after loading data (reactive loop bug)
❌ Module cards were not resizable (jqui_resizable commented out)
❌ Filter UI was a test placeholder


For the full changelog, see CHANGELOG.md.

🛠️ Development

Running in Development Mode
conda activate void R -e "shiny::runApp(port=5100)"



Running Tests
conda activate void R -e "devtools::test()"



Regenerating Documentation
conda activate void R -e "devtools::document()"



📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

🙏 AcknowledgmentsShiny — Web application framework for R
bslib — Bootstrap 5 for Shiny
dplyr — Data manipulation grammar
ggplot2 — Data visualization grammar
shinyjqui — jQuery UI interactions for Shiny
DT — Interactive data tables