Why I Built Framework

Framework exists because the same operational problems kept repeating across my projects: directory structures drifted, data sources weren’t documented, connection settings lived in stale notebooks, and every engagement started with hours of setup. Framework is the set of tools I assembled so those issues stopped consuming time.

Problems Framework Addresses

1. Project structure — every repository should have the same directories, notebooks, and defaults so handoffs take minutes, not hours.
2. Data traceability — inputs are defined once in a catalog with hashing, so I know exactly which file a notebook read.
3. Connection management — database and object storage credentials live in config plus .env, not in scripts.
4. Automation hooks — caching, git hooks, and the GUI automate repetitive tasks like syncing AI context files or running pre-commit checks.

Framework isn’t trying to be clever; it just keeps these workflows predictable.

Everyday Usage

library(framework)
project_create("my-analysis")

df <- data_read("inputs.raw.survey_2024")

results <- cache_remember("model_results", {
  train_expensive_model(data)
})

gui()

Those workflows—creating projects, registering data, caching expensive steps, and reviewing inputs/outputs in the GUI—are what Framework offers right now.

Getting Started

Install Framework from GitHub:

# install.packages("devtools")
devtools::install_github("table1/framework")

Then create your first project:

library(framework)
project_create("my-project")

Check out our documentation for detailed guides and examples.

What’s Next

I’m still building Framework and treating it as a living toolkit. If these conventions solve the same problems you run into, let me know what’s missing. The GitHub repository is the best place to:

• Report issues
• Suggest features
• Contribute code

Framework won’t eliminate every workflow wrinkle, but it removes a lot of the waste that used to slow me down. I hope it helps you do the same.