what is open science? -凯发k8网页登录

share methods and output to increase impact

open science is the practice of making scientific methodologies and output (such as publications, data, and software) transparent and broadly accessible. open science maximizes the reuse of available data and code and enables scientists to build on the work of their peers.

researchers using matlab^® and simulink^® can practice open science to increase the scientific impact of their work.

open methods: executable notebooks

well-documented methods and workflows enable open science by helping scientists follow each others’ experimental logic and interpret results.

live scripts

using matlab , researchers can tell a story with data, code, and in-line visuals to make their work process and results transparent and easy to grasp. researchers can:

• with code in executable on-line notebooks
• use interactive ui controls that allow fellow researchers to experiment with parameters and learn in a hands-on environment
• publish in standard formats such as pdf, word, html, and latex
• implement to track changes and collaborate with others on projects
• and dependency analyses to map and organize

matlab with jupyter

jupyter^® notebooks are browser-accessible computational notebooks that are popular for open science projects. as with python, matlab users can develop and run jupyter notebooks (.ipynb files) with an and a matlab license.

to interact with a multi-user jupyterhub environment, you can install the matlab integration for jupyter to provide access to a full matlab desktop via a web browser for licensed users.

open data and access

providing access to data, models, and algorithms enables scientists to build on existing frameworks. open science requires inter-operability between different data formats and programming languages, to ensure access across frameworks.

open access to data

publicly available datasets that can be accessed with matlab tools include:

• allen brain observatory: the is an object-oriented interface for accessing neuroscience data from the allen institute
• egi: users can run matlab via that can access hundreds of data centers via the egi cloud
• climate data store: ecmwf users can download netcdf files from copernicus’ c3s and import the files into matlab for processing

reading data formats

standard data and model formats also support open science. matlab can:

• work with data in popular data formats, including , and many specialized data formats, including robotics, , and time series analysis, , , and geoscience
• directly import deep learning models and
• read web application data in formats such as json, xml, and txt via matlab restful web services
• import and process data from connected devices in iot systems by combining thingspeak™ and matlab
• use arduino^® and raspberry pi™ to connect to data acquisition devices
• harness sensors on android™ or iphone devices in experiments

interoperable code and matlab algorithms

for open science, open access to data must be accompanied by open access to code and algorithms that generate meaningful results from these data.

• the bi-directional matlab integrations with python, c, fortran, and other languages enable algorithms written in one language to be called from the other
• enable sharing custom tools with non-matlab users
• with matlab coder™, users can for a variety of software and hardware platforms, including embedded systems
• matlab compiler™ enables sharing matlab applications as standalone applications, allowing end-users to run them without a matlab license

open code

a main motivation for open science is the reuse of scientific artifacts and algorithms. sharing code and software helps colleagues build on each other’s work.

open matlab code on file exchange

researchers can freely access, download, and use matlab code and algorithms shared by their peers on file exchange, or access them via the button on the matlab toolbar.

toolbox authors can share their algorithms by , ensuring a single code source with added visibility. file exchange repositories linked to github will automatically connect to the latest github version.

open matlab code on reusability portals

several reusability portals for open science offer a home for code where researchers can run uploaded code and reproduce results in the cloud, regardless of whether the researchers have a matlab license.

• : contains runnable matlab code capsules associated with research papers
• : hosts reproducible research artifacts or “tales” (matlab tales are accessible using a browser-based matlab desktop, jupyter notebook running matlab, or linux web desktop)
• : converts shared matlab code to web apps for use with data in a secure environment

open science infrastructure

as researchers increasingly connect and collaborate remotely, open science requires infrastructure beyond an individual’s local workstation.

runnable code via web browser

makes scientific resources portable and accessible on the web.

• matlab online and simulink online provide access to matlab over the web, without installation or download
• file storage on matlab drive allows researchers, educators, and students to collaborate remotely on projects and exchange artifacts

matlab on public clouds and services

matlab is accessible on various services used in open science.

• matlab works with cloud platforms, such as aws and azure, so users can scale up to take advantage of cloud compute services or access data stored in s3 containers
• data science researchers use the domino data lab platform to run matlab algorithms on centrally hosted data for scalable computing
• hpc centers such as and also offer matlab

matlab in docker containers

containerized code is portable and reusable for open science. available matlab containers include:

• a prebuilt , configurable with a built-in interactive ide
• a prebuilt on nvidia ngc cloud
• researchers can also docker containers with matlab

science gateways

science gateways are online collaboration portals where scientists and engineers access shared resources. to enable open science, mathworks and many science gateways have collaborated to make community-contributed matlab-related content and matlab compute available.

flexible licensing for more coverage

many universities and research institutes have adopted campus-wide and institute-wide matlab licensing models to provide matlab access to all their affiliated members. in keeping with open science principles, academic institutions can also provide working on a project.

open science in education

open science is not limited to research, and practices include making scientific output and methods available and accessible to students and educators. examples of publicly accessible resources for educators and students are:

• resources for with matlab via the science education resource center (serc) curriculum portal
• stem support: student competitions, partnerships, and curriculum development
• open and fun programming for children with bytes and beats courseware