Report Blocks
Reports are comprised of multiple Blocks, which wrap up all kinds of Python objects for use in building your report, such as Pandas DataFrames, Visualisations, and Markdown. We are always adding new components, and if you have some ideas on what you would like to use in your reports, please start a discussion on GitHub.
In this section we describe the Block types and provide examples. More detailed API usage can be found in our API docs.
The dp.Plot block takes a plot object from one of the supported Python visualisation libraries and renders it in your report.
Datapane currently supports the following libraries:
Library | Site / Docs |
Altair | |
Matplotlib / Seaborn | |
Bokeh | |
Plotly | |
Folium |
Altair is a declarative statistical visualization library for Python, based on Vega and Vega-Lite. Altair’s API is simple, friendly and consistent and built on top of the powerful Vega-Lite visualization grammar. This elegant simplicity produces beautiful and effective visualizations with a minimal amount of code.
To get started using Altair to make your visualizations, begin with Altair's Documentation
import altair as altimport datapane as dpfrom vega_datasets import data as vega_datagap = pd.read_json(vega_data.gapminder.url)select_year = alt.selection_single(name='select', fields=['year'], init={'year': 1955},bind=alt.binding_range(min=1955, max=2005, step=5))alt_chart = alt.Chart(gap).mark_point(filled=True).encode(alt.X('fertility', scale=alt.Scale(zero=False)),alt.Y('life_expect', scale=alt.Scale(zero=False)),alt.Size('pop:Q'),alt.Color('cluster:N'),alt.Order('pop:Q', sort='descending'),).add_selection(select_year).transform_filter(select_year)dp.Report(dp.Plot(alt_chart)).publish(name='time_interval')
Bokeh is an interactive visualization library which provides elegant, concise construction of versatile graphics, and affords high-performance interactivity over large datasets.
To get started using Bokeh to make your visualizations, begin with Bokeh's User Guide.
from bokeh.plotting import figure, output_file, showfrom bokeh.sampledata.iris import flowersimport datapane as dp# Create scatter plot with Bokehcolormap = {'setosa': 'red', 'versicolor': 'green', 'virginica': 'blue'}colors = [colormap[x] for x in flowers['species']]bokeh_chart = figure(title = "Iris Morphology")bokeh_chart.xaxis.axis_label = 'Petal Length'bokeh_chart.yaxis.axis_label = 'Petal Width'bokeh_chart.circle(flowers["petal_length"], flowers["petal_width"],color=colors, fill_alpha=0.2, size=10)# Publish the reportdp.Report(dp.Plot(bokeh_chart)).publish(name='bokeh_plot')
Matplotlib is the original Python visualisation library, often supported and used with Jupyter Notebooks. Matplotlib plots are not interactive in Datapane Reports, but are saved as SVGs so can be viewed at high fidelity.
Higher-level matplotlib libraries such as Seaborn are also supported, and can be used in a similar way to the matplotlib example below,
import matplotlib.pyplot as pltimport pandas as pdimport datapane as dpfrom vega_datasets import data as vega_datagap = pd.read_json(vega_data.gapminder.url)# gap.plot.scatter()fig = gap.plot.scatter(x='life_expect', y='fertility')dp.Report(dp.Plot(fig)).publish(name="test_mpl")
You can pass either a matplotlib Figure or Axes object to dp.Plot, you can obtain the current global figure from matplotlib by running plt.gcf()
Plotly's Python graphing library makes interactive, publication-quality graphs.
import plotly.express as pximport datapane as dpdf = px.data.gapminder()plotly_chart = px.scatter(df.query("year==2007"), x="gdpPercap", y="lifeExp",size="pop", color="continent",hover_name="country", log_x=True, size_max=60)plotly_chart.show()dp.Report(dp.Plot(plotly_chart)).publish(name='bubble')
Folium makes it easy to visualize data that’s been manipulated in Python on an interactive leaflet map. It enables both the binding of data to a map for choropleth visualizations as well as passing rich vector/raster/HTML visualizations as markers on the map.
The library has a number of built-in tilesets from OpenStreetMap, Mapbox, and Stamen, and supports custom tilesets with Mapbox or Cloudmade API keys.
If your folium map consumes live data which expires after a certain time, you can automate it to refresh the map on a cadence. See Automation.
import foliumimport datapane as dpm = folium.Map(location=[45.5236, -122.6750])dp.Report(dp.Plot(m)).publish(name='folium_map')
Markdown is a lightweight markup language that allows you to include formatted text in your report.
import datapane as dpdp.Report(dp.Markdown("__My awesome markdown__"))
To include multi-line text and formatting the words, use triple-quoted string, e.g. """Some words"""
import datapane as dpreport = dp.Report("## My Altair Plot",dp.Plot(alt_chart),dp.Markdown("""* There is a negative relation between life expectanty and fertility* The number of population with high life expactancy increases as time increase""")report.publish(name = 'results')
You can also just pass strings directly into the Report and Blocks objects, or use dp.Text, an alias of dp.Markdown
Check here for more information on how to format your text with markdown.
You can insert a Pandas dataframe in a Markdown block as nicely formatted table by using df.to_markdown(), or use the dedicated Table Block (described below) for larger data if you need sorting, filtering and more.
import pandas as pdimport datapane as dpdf = pd.read_csv("data.csv")dp.Report(dp.Markdown(df.to_markdown())
You can also share your code using dp.Markdown like below
code = dp.Markdown(f'''```pythondf = data.reset_index().melt('Date', var_name='symbol', value_name='price')base_chart = alt.Chart(df).encode(x='Date:T', y='price:Q', color='symbol').interactive()chart = base_chart.mark_line() if plot_type == 'line' else base_chart.mark_bar()```''')dp.Report(code).publish(name='report_with_code')
The Table component takes a pandas DataFrame and renders an interactive, sortable, searchable table in your Datapane report. Viewers can also download the table from the website as a CSV file.
By default, a Table only displays the first 1000 rows, and gives the user the option to load the entire dataset - datasets up to 1M cells are supported.
import datapane as dpimport pandas as pdimport numpy as npdf = pd.DataFrame({'A': np.random.normal(-1, 1, 5000),'B': np.random.normal(1, 2, 5000),})table = dp.Table(df)report = dp.Report(table)report.publish(name='sample_table')
Datapane also enables you to share files and images. For example, to share your image, you can use dp.File and pass the path.
dp.File(file='./image.jpeg')
In a similar way, any other files can be embedded in your Datapane Report, including, for instance, Excel files, JSON data, text files and more. Furthermore, Datapane will attempt to display images and JSON in your report for viewers and allow users to download them.
import datapane as dpreport = dp.Report(dp.File(file="./image.jpeg"),dp.File(file="./data.xlsx"),dp.File(file="./config.json"))report.publish(name='Files Sample')
Alternatively, instead of passing in a path, you can pass in an object such as a Python dictionary, and use is_json to tell Datapane to attempt to render it in JSON format in the front-end.
import datapane as dpthisdict = {"brand": "Ford","model": "Mustang","year": 1964}dp.Report(dp.File(thisdict, is_json=True, name='thisdict')).publish(name='json')
