Skip to content

Datapane is the easiest way to create data science reports from Python.

License

Notifications You must be signed in to change notification settings

lanthias/datapane

 
 

Repository files navigation

Datapane

Datapane Teams | Documentation | API Docs | Changelog | Twitter | Blog

Pip Downloads Latest release Conda (channel only)

Share interactive plots and data in 3 lines of Python.

Datapane is a Python library for building interactive reports for your end-users in seconds.

Import our library into your existing script/notebook and build reports from pandas Dataframes, plots from Python viz libraries, Markdown, as well as data exploration and layout components.

Export your reports as standalone HTML documents, or share and embed them via our free hosted platform.

Getting Started

Installing Datapane

The best way to install Datapane is through pip or conda.

pip

$ pip3 install -U datapane
$ datapane hello-world

conda

$ conda install -c conda-forge "datapane>=0.12.0"
$ datapane hello-world

Datapane also works well in hosted Jupyter environments such as Colab or Binder, where you can install as follows:

!pip3 install --quiet datapane
!datapane signup

Explainer Video

Datapane.Public.Tutorial.mp4

Hello world

Let's say you wanted to create a report with an interactive plot and table viewer:

import altair as alt
from vega_datasets import data
import datapane as dp

source = data.cars()

plot1 = alt.Chart(source).mark_circle(size=60).encode(
  x='Horsepower',
  y='Miles_per_Gallon',
  color='Origin',
  tooltip=['Name', 'Origin', 'Horsepower', 'Miles_per_Gallon']
).interactive()

dp.Report(
    dp.Text("## Hello world!"),
    dp.Plot(plot1),
    dp.DataTable(source)
).save(path="Hello_world.html")

This will package a standalone HTML document that looks as follows:

Simple Datapane report example with text, plot and table

Your users can scroll & zoom on the chart, filter and download the tabular data.

Advanced Layout Options

Datapane is great for presenting complex data and provides many components for creating advanced interactive layouts. Let's you need to write a technical document:

import altair as alt
from vega_datasets import data
import datapane as dp

source = data.cars()
plot1 = alt.Chart(source).mark_circle(size=60).encode(
    x='Horsepower',
    y='Miles_per_Gallon',
    color='Origin',
    tooltip=['Name', 'Origin', 'Horsepower', 'Miles_per_Gallon']
).interactive()

dp.Report(
    dp.Page(title="Charts and analysis",
            blocks=[
                dp.Formula("x^2 + y^2 = z^2"),
                dp.Group(
                    dp.BigNumber(
                        heading="Number of percentage points",
                        value="84%",
                        change="2%",
                        is_upward_change=True
                    ),
                    dp.BigNumber(
                        heading="Simple Statistic",
                        value=100
                    ), columns=2,
                ),
                dp.Select(blocks=[
                    dp.Plot(plot1, label="Plot"),
                    dp.HTML('''<iframe width="560" height="315" src="https://www.youtube.com/embed/dQw4w9WgXcQ" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>''', label="Video")
                ]),
            ]),
    dp.Page(title="Dataset", blocks=[
            dp.DataTable(source)
    ])
).save(path="Complex_layout.html", open=True)

Layout blocks like dp.Select, dp.Group and dp.Page allow you to highlight key points without sacrificing detail, while content blocks like dp.HTML and dp.Formula (LaTeX) can enrich your report. The final result looks like this:

Complex Datapane report example

Check out the full list of blocks in our documentation.

Sharing Reports

Sign up for a free account

In addition to saving documents locally, you can host, share and embed reports via Datapane Studio.

To get your free API key, run the following command in your terminal to sign up via email/OAuth:

$ datapane signup

If you're using Jupyter, run !datapane signup instead.

Next, in your Python notebook or script simply change the save function to upload on your report:

dp.Report(
 ...
#).save(path="hello_world.html")
).upload(name="Hello world")

Your Studio account comes with the following:

  • Unlimited public reports - great for embedding into places like Medium, Reddit, or your own website (see here)
  • 5 private reports - share these via email within your organization

Featured Examples

Here a few samples of the top reports created by the Datapane community. To see more, check out our gallery section.

Teams

Datapane Teams is our plan for teams, which adds the following features on top of our open-source and Studio plans:

  • Private domain and organizational workspace
  • Multiple projects
  • Client-sharing functionality
  • Unlimited Datapane Apps
  • Custom App packages and environments
  • Secure Warehouse & API Integration
  • File and Dataset APIs
  • Private Slack or Teams support

Datapane Teams is offered as both a managed SaaS service and an on-prem install. For more information, see the documentation. You can find pricing here.

Next Steps

Analytics

By default, the Datapane Python library collects error reports and usage telemetry. This is used by us to help make the product better and to fix bugs. If you would like to disable this, simply create a file called no_analytics in your datapane config directory, e.g.

Linux

$ mkdir -p ~/.config/datapane && touch ~/.config/datapane/no_analytics

macOS

$ mkdir -p ~/Library/Application\ Data/datapane && touch ~/Library/Application\ Data/no_analytics

Windows (PowerShell)

PS> mkdir ~/AppData/Roaming/datapane -ea 0
PS> ni ~/AppData/Roaming/datapane/no_analytics -ea 0

You may need to try ~/AppData/Local instead of ~/AppData/Roaming on certain Windows configurations depending on the type of your user-account.

Joining the community

Looking to get answers to questions or engage with us and the wider community? Check out our GitHub Discussions board.

Submit feature requests, issues, and bug reports on this GitHub repo.

Open-source, not open-contribution

Datapane is currently closed to external code contributions. However, we are tremendously grateful to the community for any feature requests, ideas, discussions, and bug reports.

About

Datapane is the easiest way to create data science reports from Python.

Resources

License

Code of conduct

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 86.3%
  • Jupyter Notebook 12.8%
  • Other 0.9%