DVC and Git For Data Science

Explore what git is, a distributed version control system that tracks changes across working, staging, local, and remote repositories, enabling seamless collaboration and reliable code history.

Master the basic git workflow: initialize, stage changes with git add, commit locally, and push to a remote repository for collaboration on GitHub, Bitbucket, or GitLab.

Learn to configure git with global settings by setting user name and email via git config, then verify and use these credentials for pushes across all repositories.

Explore using GitHub as a remote code repository, including signing up, creating repositories, cloning and pushing code, and managing issues and pull requests.

generate an ssh key pair, optionally set a passphrase, and add the public key to GitHub settings to enable push access without passwords.

Create a new repository on GitHub, initialize with a README, and clone it locally; then stage changes, commit, and push to the remote to track versions.

Learn a practical git workflow for data science projects using GitHub: initialize a repo, stage changes, commit with clear messages, and push to update remote repositories.

Learn to write effective git commit messages by recording changes, preserving history, and using a subject and body with an imperative, capitalized subject, and a blank line between them.

Learn how to undo uncommitted changes in git by discarding edits in the staging area with git checkout -- and by using git restore or git reset to undo commits.

Learn to use git log to view and filter repository history, inspect commit hashes, dates, and messages, and search logs by time, branch, or path with practical examples.

Explore branch types: main branch, master branch, and develop branch, plus feature, release, and hot phase rate branches; learn to create, switch, rename, and delete branches locally and remotely.

Learn practical git basics for data science workflows, including tracking files, identifying ignored vs untracked files, and essential commands such as init, clone, pull, push, and fetch to manage changes.

Master advanced GitHub search to find code, files, and readmes across organizations and users using language, location, stars, and combined filters. Includes searching readmes, descriptions, and code within repositories.

Explore GitHub's advanced search features, learning to filter repositories by name, description, read me, language, user or organization, location, and activity, and to combine qualifiers for precise results.

Design a CLI for a Git-like version control workflow, covering init, add, commit, log, and push to a remote repository, with Python and Click-based command-line patterns.

Build a version control cli status function to track file changes by comparing current and previous states using modification times. Log added, removed, and modified files with dictionary diffs.

Implement push functionality in a version control CLI for data science, covering pushing from local to remote repositories, prompts for credentials, and configuring the remote origin.

Learn to build a version control CLI for data science workflows by implementing remove, restore, and clone operations, including repository cloning, commit history tracking, and staging changes.

DVC provides data version control for data science, tracking changes across versions of data and models in a history alongside code, with on premise or cloud storage and DVC Studio.

Explore the features and benefits of DVT, including versioning for data, experiment tracking, and building powerful pipelines to streamline machine learning workflows.

Master the three DVC areas—working area, staging area, and cache—by initializing with dvc init, adding data and models with dvc add, and pushing to a remote repository with dvc push.

Data version control tracks changes to data and code and lets you revert to prior versions. It acts as a single source of truth across teams and speeds up collaboration.

this lecture compares dvc and git commands, showing their similarities and easy adoption. it covers dvc init, get, status, add, commit, push, and input, alongside corresponding git commands.

Learn how to use DVC in data science to version and track data, models, and assets alongside Git. Follow the workflow from initialization to pushing to remote storage.

Learn to push data with DVC to Google Drive by configuring a DVC remote, installing the drive plugin, and executing dvc push to store real data remotely.

Initialize the repository, add a local remote storage, and push data to the dvc project's remote. Use pull or check out to restore data from the remote when needed.

Explore the DAGsHub platform walkthrough to manage data science projects with experiment tracking, model comparisons, DVC integration, ten gig free storage, and a GitHub-like interface.

Create and customize a new dagsHub repository by signing in, choosing public or private visibility, adding a description and readme, selecting a license, and initializing with remote or local setup.

Discover how to search on dagshub for topics, users, and organizations, fetch repositories, and explore topic-based organization for data science projects.

Learn how to add topics on DAGsHub to improve repository discoverability by selecting predefined or general topics, adding multiple topics, submitting changes, and deleting topics as needed.

Learn how to use DAGsHub with Label Studio to create a medical annotation project: upload data, configure labels (drug, dosage, disease, symptoms, procedure, treatment), and commit annotated results to GitHub.

Develop an end-to-end data science workflow by using Git and DVC to manage data, structure projects, engineer features, train and interpret models, and deploy with fast api and GitHub tracking.

Learn to version and push data with DVC and Git, initializing the workspace, tracking data, committing changes, and pushing to Dagshub for efficient data collaboration.

Prepare data with pandas, apply label encoding for gender, and build a workflow pipeline to train and evaluate classification models, interpret accuracy and precision.

Learn practical git branching for machine learning workflows: create and manage branches (main, develop, feature, release, hotfix), switch between them, rename, view, and delete branches for model versions.

Learn to store a model with git and dvc, push to DagsHub via development and master branches, and manage versions using dvc push and origin commands.

Learn to save a new ML model to a new git branch using DVC for tracking, commit changes, and push to remote branches for model versioning.

Learn to track machine learning experiments with DagsHub, logging metrics via yaml configurations, recording model parameters, comparing base and improved models, and pushing results through git workflows.

Explore manual ml pipelines by building datasets, processing data with pandas, encoding features, and training a model, then evaluate with metrics and scores to compare performance.

Learn to automate data and model tracking with DVC pipelines using DVC run interactive experiments, define dependencies and outputs, and visualize the workflow with a dag.

Create and manage dvc pipelines by defining stages, dependencies, and outputs; run data preparation, model training, and evaluation steps with dvc run to reproduce experiments.

Learn how to use DVC metrics to extract, store, and explore metrics from a data science pipeline, compare runs, and visualize results with markdown tables.

Explore using DVC pipelines to reproduce data science workflows, commit changes, manage training data and models, and compare training results and metrics across experiments.

Push your DVC pipelines from your workspace to DagsHub by configuring a remote and private repository, then push data and code with git and DVC.

Learn to fix push and pull errors in DVC pipelines by performing a force push to the remote, verify the push, and confirm branch updates.