ResearchCompendia is a web service allowing people to share the research software and data associated with a scientific publication (articles and working papers).
With ResearchCompendia we provide the tools to publish digital scholarly objects by hosting data, code, and methods in a form that is accessible, trackable, and persistent. It has been a long held principle of academic science that all respectable research must be reproducible. Now that much of that data and analysis is generated computationally, it is natural to expect that code and parameters also be shared.
With ResearchCompendia we wish to address these issues facing reproducibility in computational research. We will provide tools to share and archive the data, codes, documentation, parameters, and environmental settings linked with published research all in one place. Soon, we wish to support the verification and validation processes by providing for the remote execution of shared codes in our cloud resources, and the visualization of results. Most of all we wish to make these tools heavily automated, and easy to access and utilize to lessen the exertion required from already overburdened academic researchers in the process of publishing fully reproducible work.
The ResearchCompendia website is operated not for profit. The mission is to make research easier to use and easier to replicate. ResearchCompendia is currently funded by several national research agencies and universities.
We introduce the concept of a compendium as both a container for the different elements that make up the document and its computations (i.e. text, code, data,...), and as a means for distributing, managing and updating the collection.
Gentleman, R. and Temple Lang, D. "Statistical Analyses and Reproducible Research" (May 2004). Bioconductor Project Working Papers.
A Research Compendium is a digital scholarly object comprising all relevant narrative, code, and data required to present and disseminate published research including computational methods as truly reproducible science. So here we provide the tools to easily create such a compendium by allowing you to easily and elegantly link your publications to your data, codes, and detailed computational methods information and parameters, and any further materials necessary to allow other researchers to reproduce your computational procedures.
You can browse existing compendia here.
ResearchCompendia follows three main objectives:
To do this we hope to provide the following tools:
ResearchCompendia is intended both as a tool to improve the communication of science, and to make the lives of publishing scientists easier. There are many reasons to create a compendium with our site:
Click Create Compendia to start your own.
Depending on your interest there are several ways you can get involved:
Research without proper citation and attribution is not reproducible research. Thus, proper attribution is a priority for ResearchCompendia.
Each and every Compendium should have a Cite Me! link that will provide you with multiple formats by which to cite the research, data and code on that page.
We will shortly be adding DOI assignments options for data and code contributions.
You can upload or link any kind of code or data related to published scientific research. When we have remote execution available, we will start with the languages most in demand by our existing users, and languages for which easy APIs and other tools exists (starting with R, MatLab, Python, and Cactus). While this limited but ever expanding language set will be all that is executable on a Compendium page, any type of code may be hosted or linked through us as downloadable content.
We do not yet have specific code and data size limitations, but try to keep uploads within reason (<20 GB).
We have an easy to use form for page creation. You can much of the metadata about your paper is automatically filled in if your publication has a DOI. We offer many ways of linking or uploading your code and data to us.
The time required for page creation depends upon how much detail you want to add into the site itself about parameter usage and computational methodology. Page creation can take as little at 10 minutes.
Click Create Compendia to start your own.
ResearchCompendia does not take a copyright for any computer codes, data, or other documents researchers post on the ResearchCompendia website.
We will track all changes to compendia pages, and changes will always be noted on the site. Obviously, persistency is essential to making ResearchCompendia effective. We are currently working towards a sustainable funding plan to ensure that all sites stay up and active for the long term. However, independent of our funding, should anything happen to ResearchCompendia, we will transfer all data and code to appropriate hosts (such as our partners like Github or Dataverse) to ensure the persistence of your information and your DOIs.
We are independent from any of the publishing companies or scientific journals, but are working to support the movement toward more direct access to journals as a tool for the sharing of supplementary materials for the promotion of reproducible research and transparency in publications.
Yes! We started with applications in economics and statistics, but the mission of ResearchCompendia is universal across the sciences and the framework can be extended to any discipline. We plan to cover astrophysics, computational neuroscience, computational biology, and we eventually hope to have contributions from all fields of science using original computational methods! If you have a request for a particular functionality that would be useful to you or your subfield please tell us!
All services provided by our site are free to use. We provide hosting and computational cloud resources for all researchers. We are a not-for profit organization. We are, of course, always looking for sustainable funding sources to cover our expenses, so if you are interested in contributing, please contact us.
The creation of a ResearchCompendia account is completely free of charge. First, click on the register link on the top right corner of the webpage. Fill in the form with your first name, last name, and a valid email address. You will automatically receive an email with a URL that will allow you to complete the registration process.
You have unlimited access to all the companion websites. Once you have logged in, you can create and manage your own companion websites to promote your own research.