You may have had the need to visualize data in a more graceful and organic way than using pandas commands (like df.plot and, for seaborn, the excellent sns.pairplot) maybe simply because you need a web interface (even if it is a local server) to which a third-party non-specialist user can access and you do not want to resort to fastapi.
Then think about Dash and Streamlit.
Both Dash and Streamlit are designed to facilitate the creation of web applications dedicated to data science, but they differ in some characteristics:
Abstraction:
Dash: Provides a higher level of abstraction, allowing you to create complex and interactive layouts using a language that resembles HTML, CSS and JavaScript. This gives you more control over the user interface, but requires a bit more time to learn.
Streamlit: It has a simpler and more direct API, focused on fast data visualization. It is perfect for creating prototypes or basic dashboards quickly.
Components:
Dash: Includes a wide range of ready-made components, such as charts, maps and tables, although custom components may need to be developed for specific needs.
Streamlit: Has fewer components than Dash, but the offering is growing. Its simplicity makes it suitable for many common scenarios.
Layout:
Dash: Uses a traditional grid and container approach to layout, allowing for more control over elements, but requiring more code.
Streamlit: Uses a simpler, more declarative method, where elements are added sequentially, making the process quick and intuitive.
Community and Ecosystem:
Dash: Has a larger, more established community, with many resources and tutorials available.
Streamlit: Its community is younger but growing.
When to use one or the other?
Dash: It is the best choice if you need highly customized and interactive dashboards, with fine-grained control over the user interface. Ideal for complex projects and teams with knowledge of web technologies.
Streamlit: Perfect if you want to create simple prototypes or dashboards quickly, especially for data scientists who want to share results without delving into web development.
Bottom line: Both tools are useful for data visualization, but choosing between Dash and Streamlit depends on your needs and skills. If you are looking for flexibility and control, Dash is a better fit; if you are looking for simplicity and speed, Streamlit is for you.
Additional considerations:
Integration with other libraries: Both can be integrated with libraries such as Pandas, NumPy and scikit-learn.
Deployment: Dash and Streamlit can be deployed on platforms such as Heroku, AWS and others.
Examples
It is very easy to find code already written using these 2 libraries: an example by Amanda Kieswetter
or this tutorial roduction to Dash Plotly – Data Visualization in Python by Charming Data.