Visualize 100 million points with Datashader and Plotly

Continuum Analytics has a super new open-source project called “Datashader.” You can download it here through their Anaconda or Miniconda packages:

https://www.continuum.io/downloads

Or you can install it via pip like so:

Datashader can plot huge amounts of data and display it as a static image. In this post, we plot 100 million points using Datashader, then convert the result to an interactive graph using Plotly’s Python library. Datashader employ min-max decimation in Python to get a zoomed-out, visual view of “big data.” It’s an open-source project, so check it out!

Get the full Jupyter notebook for this post online at Plotly.

1. IMPORTS

2. GENERATE FAKE TIME SERIES DATA

Use floating point UNIX time instead of datetimes for speed.

Modified slightly from: https://anaconda.org/jbednar/tseries/notebook

3. GENERATE AN IMAGE OF THE DATA USING DATASHADER

téléchargement

Nice, an image of our 100 million point time series!

4. CONVERT THE IMAGE TO AN INTERACTIVE CHART USING PLOTLY

 

Get the full Jupyter notebook for this post online at Plotly.