Open Science with MATLAB
Researchers can use MATLAB® for their Open Science projects and make their scientific output available and accessible for their communities.
MATLAB supports a variety of data formats and can interface with publicly available databases, allowing researchers easy access to data and the ability to analyze open data.
Creating computational notebooks with MATLAB live scripts produces readable code and helps create a story combining code, images, results, and equations all in one document.
MATLAB File Exchange contains more than 40,000 open-source MATLAB code repositories developed by users that are freely available either via the web platform or as add-ons from the MATLAB desktop interface. Researchers looking to make their MATLAB code reusable and reproducible can now upload them on Open Science reproducibility platforms like Code Ocean and WholeTale. These sites provide a permanent home for scientific code where it can be run and tested by everybody, including those without a license. MATLAB interoperability with other languages including Python® enables hybrid workflows where researchers can combine several languages for complex projects.
Finally, when it comes to scaling up beyond local resources, MATLAB cloud offerings let researchers use their licenses with remote HPC centers, Docker containers, science gateways, and cloud compute services such as AWS® and Azure®.
Published: 12 Sep 2021
Welcome to this short video on open science with MATLAB, me Shubo. How does open science actually help research? Open science makes scientific resources freely available. Now, you probably care about this if you are either a scientist or an engineer using or contributing to such resources, or a system administrator or science gateway owner building or supporting sharing infrastructures.
Let's start with the users. How can MATLAB help with your reusable research needs? Creating and accessing data. Good, clean data takes effort to collect. Today, there are several public databases making standardized, curated data available according to fair standards. One example is the EGI, a European federated network that can connect you to hundreds of data centers via a centralized cloud. Having a MATLAB interface to such data sets with support for a wide range of data formats helps you access data from MATLAB.
Producing readable code. Instead of rewriting every single analysis ever done, wouldn't it be great to simply reuse code from others and build upon them? For this, code needs to be readable and organized. Computational notebooks using MATLAB Live Scripts helps you paint a story combining your code with images, text, and inline output. Sharing research output. Most research output gets lost after publication. How can you increase the discoverability of your research? Uploading your code onto multiple listings and platforms helps your peers access your code. Interoperability with other languages means your MATLAB code is also accessible by those not using MATLAB.
Reproducibility and reusability are big issues for open science. Today, several code sharing platforms are used by researchers to exchange software. MATLAB File Exchange offers more than 40,000 freely available widely used MATLAB repositories. Toolmakers can now directly link to their GitHub contributions for added visibility. Reusability platforms like Code Ocean and Whole Tale provide a permanent home for your published MATLAB code where everyone, even those without licenses, can run and test it freely via a browser.
Now, sometimes your research needs resources beyond your local workstation. For this, MATLAB users can now scale up and make use of public clouds and other resources. For example, if you need parallel computing on clusters to speed up your code, you can use your campus-wide license to access all available workers on your cluster, either locally or at a remote HPC center. If you're training a deep learning model, use the MATLAB Deep Learning Container on the NVIDIA NGC Cloud to speed it up. Finally, take advantage of cloud compute on AWS and Azure by using your MATLAB code on these services.
As you've seen, open science resource sharing can involve many components. Now, whether you are a scientist or an engineer using these shared resources or a system administrator or science gateway owner supporting open science infrastructures, you can pick and choose the components for your workflow all from within the MATLAB environment. So go ahead and share your open MATLAB code with your community, or get in touch if you want your science gateway to partner with MathWorks. Thank you.