Skip to content

Latest commit

 

History

History
60 lines (41 loc) · 2.26 KB

notes-vistrails.md

File metadata and controls

60 lines (41 loc) · 2.26 KB

Where to find it?

https://www.vistrails.org/index.php/Main_Page

What is it?

An open-source scientific workflow and provenance management system that supports data exploration and visualization.

Runs on Mac, windows and linux. Written in Python and uses the multi-platform QT library.

Comes with pylab, matplotlib, imagemagick, and other packages for visualization.

For exploratory science like simulations, data analysis and visualization, the workflow is dynamically created and evolves rapidly. VisTrails manages the workflow.

Example: latex figure

Say you run a simulation and make a figure using VisTrails and embed it in your latex document (your paper). Then a person reading your paper (electronically) can click on the figure to obtain the VisTrails workflow by clicking on it. This allows them to reproduce your simulation and recreate your figure!

Who uses it?

  • NASA in their climate data analysis tool. See I tried to get the data and reproduce it, but it requires a little bit of work.
  • NSF Center for Coastal Margin Observation and Prediction
  • USGS Habitat Modeling.

Installation

sudo pip install vistrails PyQt4 tej scikit-learn

Or download a windows or mac installer and follow the instructions on the website.

Demonstration

You construct a workflow in vt file. I looked at weather.vt in the examples directory. Open it in the gui. You can see several plots there. There is a workflow for a "Temperature Histogram". It will display a flowchart that does the following:

  1. It first downloads a file with the data. The url string there if you click on it says

    http://www.vistrails.org/download/download.php?type=DATA&id=weather_data.zip

  2. Then it unzips the file data.zip.

  3. The next field describes a csv file that ought to have been unzipped.

  4. Next, it extracts a column from the csv file called GetTemperature from this file.

  5. It forms a histogram using matplotlib in the usual way.

  6. Then it formats the figure, adds axes properties and sets a legend.

  7. It writes the figure to the so-called vistrails "spreadsheet"

Now, hit execute in your vistrails window. You should see a graph of a temperature histogram! This is reproducible research, isn't it guys?!