We're really excited to announce the release of our newest product, Yhat ScienceBox!
In working on data science projects, we've often found that spinning up new servers, ssh-ing into boxes, and waiting for lengthy installs of Python and R libraries can be a real pain. For people new to data science or data scientists who simply aren't as technically oriented, offloading computationally intensive jobs to external servers can be a daunting task.
ScienceBox is a standalone analytics server expressly designed for doing data science.
The Ubuntu-based application stack ships with hundreds of well-known packages for R and Python, RStudio Server and IPython Notebook, and a dead-simple package and environment management system.
(escape from dependency hell)
ScienceBox let's you manage packages and dependencies via the browser. Pop open
/packages page and install all your favorite Python, R, and command line
Work in your browser
Need a little extra horsepower for your R or Python scripts? ScienceBox ships with RStudio and IPython Notebook configured, so you can run all of your analysis remotely. Gone are the days of keeping your laptop running overnight!
Command line goodness (the
We know many people (ourselves included) prefer working on data science projects in a text editor or from command line.
ScienceBox comes with a command line client,
sb, that let's you interact with
a remote ScienceBox via your laptop. Just prefix your regular commands with
Manage your files
$ sb sync
sb handle transferring your scripts and data from your laptop to the server
. Stop doing the
python run.py dance!
And everything is there, ready and waiting for your in on the
And yes, it does syntax highlighting!
Offload big scripts
Jobs can be run directly from command line:
Want to unleash the power of 8 cores and 68 GB of RAM? Use
sb run to quickly
and easily execute server-side commands.
$ sb run python sklearn_demo.py
You can always find your results on the Jobs page:
Don't have time to sit and wait for your job to finish? Let your ScienceBox send you an email when it's done.
$ sb run -e email@example.com python sklearn_demo.py
Dead simple scheduling
And of course for those jobs you want to re-run, ScienceBox ships with a simple scheduler.
Hope this gives you a better feel for ScienceBox. If you have any feedback, feature requests or any suggestions for the product, please email us and we'll try to incorporate your input into the next release.
Want your own ScienceBox?
Head over to the AWS Marketplace
For more documentation on ScienceBox: ScienceBox Documentation