Cookiecutter Data Science: How to Organize Your Data Science Project

A logical, reasonably standardized, but flexible project structure for doing and sharing
c
comments

By DrivenData

Image

Why use this project structure?

We're not talking about bikeshedding the indentation aesthetics or pedantic formatting standards — ultimately,

cookiecutter https://github.com/drivendata/cookiecutter-data-science


Example

 

Directory structure

├── LICENSE
├── Makefile           <- Makefile with commands like `make data` or `make train`
├── README.md          <- The top-level README for developers using this project.
├── data
│   ├── external       <- data from third party sources.
│   ├── interim        <- Intermediate data that has been transformed.
│   ├── processed      <- The final, canonical data sets for modeling.
│   └── raw            <- The original, immutable data dump.
│
├── docs               <- A default Sphinx project; see sphinx-doc.org for details
│
├── models             <- Trained and serialized models, model predictions, or model summaries
│
├── notebooks          <- Jupyter notebooks. Naming convention is a number (for ordering),
│                         the creator's initials, and a short `-` delimited description, e.g.
│                         `1.0-jqp-initial-data-exploration`.
│
├── references         <- data dictionaries, manuals, and all other explanatory materials.
│
├── reports            <- Generated analysis as HTML, PDF, LaTeX, etc.
│   └── figures        <- Generated graphics and figures to be used in reporting
│
├── requirements.txt   <- The requirements file for reproducing the analysis environment, e.g.
│                         generated with `pip freeze > requirements.txt`
│
├── setup.py           <- Make this project pip installable with `pip install -e`
├── src                <- Source code for use in this project.
│   ├── __init__.py    <- Makes src a Python module
│   │
│   ├── data           <- Scripts to download or generate data
│   │   └── make_dataset.py
│   │
│   ├── features       <- Scripts to turn raw data into features for modeling
│   │   └── build_features.py
│   │
│   ├── models         <- Scripts to train models and then use trained models to make
│   │   │                 predictions
│   │   ├── predict_model.py
│   │   └── train_model.py
│   │
│   └── visualization  <- Scripts to create exploratory and results oriented visualizations
│       └── visualize.py
│
└── tox.ini            <- tox file with settings for running tox; see tox.testrun.org


Opinions

 
There are some opinions implicit in the project structure that have grown out of our experience with what works and what doesn't when collaborating on data science projects. Some of the opinions are about workflows, and some of the opinions are about tools that make life easier. Here are some of the beliefs which this project is built on—if you've got thoughts, please contribute or share them.

Data is immutable

 
Don't ever edit your raw data, especially not manually, and especially not in Excel. Don't overwrite your raw data. Don't save multiple versions of the raw data. Treat the data (and its format) as immutable. The code you write should move the raw data through a pipeline to your final analysis. You shouldn't have to run all of the steps every time you want to make a new figure (see Analysis is a DAG), but anyone should be able to reproduce the final products with only the code in src and the data in data/raw.

Also, if data is immutable, it doesn't need source control in the same way that code does. Therefore, by default, the data folder is included in the .gitignore file. If you have a small amount of data that rarely changes, you may want to include the data in the repository. Github currently warns if files are over 50MB and rejects files over 100MB. Some other options for storing/syncing large data include AWS S3 with a syncing tool (e.g., s3cmd), Git Large File Storage, Git Annex, and dat. Currently by default, we ask for an S3 bucket and use AWS CLI to sync data in the data folder with the server.

Notebooks are for exploration and communication

 
Notebook packages like the Jupyter notebook, Beaker notebook, Zeppelin, and other literate programming tools are very effective for exploratory data analysis. However, these tools can be less effective for reproducing an analysis. When we use notebooks in our work, we often subdivide the notebooks folder. For example, notebooks/exploratory contains initial explorations, whereas notebooks/reports is more polished work that can be exported as html to the reports directory.

Since notebooks are challenging objects for source control (e.g., diffs of the json are often not human-readable and merging is near impossible), we recommended not collaborating directly with others on Jupyter notebooks. There are two steps we recommend for using notebooks effectively:

  1. Follow a naming convention that shows the owner and the order the analysis was done in. We use the format <step>-<ghuser>-<description>.ipynb (e.g., 0.3-bull-visualize-distributions.ipynb).
  2. Refactor the good parts. Don't write code to do the same task in multiple notebooks. If it's a data preprocessing task, put it in the pipeline at src/data/make_dataset.py and load data from data/interim. If it's useful utility code, refactor it to src.

Now by default we turn the project into a Python package (see the setup.py file). You can import your code and use it in notebooks with a cell like the following:

# OPTIONAL: Load the "autoreload" extension so that code can change
%load_ext autoreload

# OPTIONAL: always reload modules so that as you change code in src, it gets loaded
%autoreload 2

from src.data import make_dataset


Analysis is a DAG

 
Often in an analysis you have long-running steps that preprocess data or train models. If these steps have been run already (and you have stored the output somewhere like the data/interim directory), you don't want to wait to rerun them every time. We prefer make for managing steps that depend on each other, especially the long-running ones. Make is a common tool on Unix-based platforms (and is available for Windows). Following the make documentation, Makefile conventions, and portability guide will help ensure your Makefiles work effectively across systems. Here are some examples to get started. A number of data folks use make as their tool of choice, including Mike Bostock.

There are other tools for managing DAGs that are written in Python instead of a DSL (e.g., Paver, Luigi, Airflow, Snakemake, Ruffus, or Joblib). Feel free to use these if they are more appropriate for your analysis.

Build from the environment up

 
The first step in reproducing an analysis is always reproducing the computational environment it was run in. You need the same tools, the same libraries, and the same versions to make everything play nicely together.

One effective approach to this is use virtualenv (we recommend virtualenvwrapper for managing virtualenvs). By listing all of your requirements in the repository (we include a requirements.txt file) you can easily track the packages needed to recreate the analysis. Here is a good workflow:

  1. Run mkvirtualenv when creating a new project
  2. pip install the packages that your analysis needs
  3. Run pip freeze > requirements.txt to pin the exact package versions used to recreate the analysis
  4. If you find you need to install another package, run pip freeze > requirements.txt again and commit the changes to version control.

If you have more complex requirements for recreating your environment, consider a virtual machine based approach such as Docker or Vagrant. Both of these tools use text-based formats (Dockerfile and Vagrantfile respectively) you can easily add to source control to describe how to create a virtual machine with the requirements you need.

Keep secrets and configuration out of version control

 
You really don't want to leak your AWS secret key or Postgres username and password on Github. Enough said — see the Twelve Factor App principles on this point. Here's one way to do this:

Store your secrets and config variables in a special file

Create a .env file in the project root folder. Thanks to the .gitignore, this file should never get committed into the version control repository. Here's an example:

# example .env file
DATABASE_URL=postgres://username:password@localhost:5432/dbname
AWS_ACCESS_KEY=myaccesskey
AWS_SECRET_ACCESS_KEY=mysecretkey
OTHER_VARIABLE=something


Use a package to load these variables automatically.

If you look at the stub script in src/data/make_dataset.py, it uses a package called python-dotenv to load up all the entries in this file as environment variables so they are accessible with os.environ.get. Here's an example snippet adapted from the python-dotenv documentation:

# src/data/dotenv_example.py
import os
from dotenv import load_dotenv, find_dotenv

# find .env automagically by walking up directories until it's found
dotenv_path = find_dotenv()

# load up the entries as environment variables
load_dotenv(dotenv_path)

database_url = os.environ.get("DATABASE_URL")
other_variable = os.environ.get("OTHER_VARIABLE")


AWS CLI configuration

When using Amazon S3 to store data, a simple method of managing AWS access is to set your access keys to environment variables. However, managing mutiple sets of keys on a single machine (e.g. when working on multiple projects) it is best to use a credentials file, typically located in ~/.aws/credentials. A typical file might look like:

[default]
aws_access_key_id=myaccesskey
aws_secret_access_key=mysecretkey

[another_project]
aws_access_key_id=myprojectaccesskey
aws_secret_access_key=myprojectsecretkey


You can add the profile name when initialising a project; assuming no applicable environment variables are set, the profile credentials will be used be default.

Be conservative in changing the default folder structure

 
To keep this structure broadly applicable for many different kinds of projects, we think the best approach is to be liberal in changing the folders around for your project, but be conservative in changing the default structure for all projects.

We've created a folder-layout label specifically for issues proposing to add, subtract, rename, or move folders around. More generally, we've also created a needs-discussion label for issues that should have some careful discussion and broad support before being implemented.

Contributing

 
The Cookiecutter data Science project is opinionated, but not afraid to be wrong. Best practices change, tools evolve, and lessons are learned. The goal of this project is to make it easier to start, structure, and share an analysis. Pull requests and filing issues is encouraged. We'd love to hear what works for you, and what doesn't.

If you use the Cookiecutter data Science project, link back to this page or give us a holler and let us know!

Links to related projects and references

 
Project structure and reproducibility is talked about more in the R research community. Here are some projects and blog posts if you're working in R that may help you out.

  • Project Template - An R data analysis template
  • "Designing projects" on Nice R Code
  • "My research workflow" on Carlboettifer.info
  • "A Quick Guide to Organizing Computational Biology Projects" in PLOS Computational Biology

Finally, a huge thanks to the Cookiecutter project (github), which is helping us all spend less time thinking about and writing boilerplate and more time getting things done.

 
Bio: DrivenData is a mission-driven data science firm that brings the powerful capabilities of data science, machine learning, and artificial intelligence to organizations tackling the world’s biggest challenges. DrivenData Labs (drivendata.co) helps mission-driven organizations harness data to work smarter, offer more impactful services, and use machine intelligence to its fullest potential. DrivenData also runs online machine learning competitions (drivendata.org) where a passionate, global community of data scientists build algorithms for social impact.

Original. Reposted with permission.