A Researcher's guide on Open Data, Open-source Software, and Open AI

Here at the Open@UTA blog series, we have talked about Openbefore, which, as defined by the Open Knowledge Foundation, is Knowledge that is freely accessible to anyone, useable, modifiable, and re-shareable. The concept of Open Data is similar; as defined in the Open Data Handbook, it is data that is freely available, accessible, useable, re-useable, and redistributable.

Opensource.com defines any software with free access to the source code and the ability to inspect, modify, and enhance it as Open-source software. While the definition of Open AI (as a concept and not the organization) is not as common so far, the idea Jennifer Ding talks about is Artificial Intelligence (AI) systems and models that allow for the complete and free use, re-use, collaboration, distribution, and participation in the algorithm and model development.

Although the subject matter may change, the concept is clear, and there is a need for "Openness." Without open data and open software, many students and researchers lose the opportunity to conduct free research and must pay to access and, in return, produce more resources. Open-source software has benefited greatly from free-lance software programmers that enhance the software's quality, fairness, and efficiency and allow novice programmers to learn and share their skills online. Similarly, Open AI initiatives, such as offering open-source large language models  (LLMs), are especially useful for researchers, start-ups, and businesses.

While I agree there is more that can be done, initiatives such as the OSTP Nelson Memo have already been taken to ensure that federally funded publications and supporting data are freely available and accessible without an embargo.

Where to look?

There are many good resources for finding data, such as the Texas Open Data Portal, the Open Data Network, Kaggle, UN Data,  Datahub, and UTA Libraries' open-access MavMatrix.

You can find a list of open data sources classified by the subject created by Pavlo Rymarchuk at Kaggle.

Find a list of Open-source software alternatives to proprietary software at opensourcealternative.to and a list of open-source LLMs for 2024 as shared by the Elastic Platform Team.

Sharing your work Openly?

Many platforms, including the UTA Library's MavMatrix, support the ability to share and publish your work openly. Amazon's AWS Open Data Sponsorship Program allows you to store your publicly available high-valued cloud-optimized datasets. Amazon covers the data storage and data transfer costs for up to two years. Google's Open-Source platform is also a place you may want to explore for your Open-source projects.

Support for Open at UTA Libraries

For researchers and faculty interested in incorporating Open in their coursework, the Libraries offer support with Open Education Grants (OER) and Research data support at the Day Family Research Lab (DFRL). The annual library event UTA Datathon offers data enthusiasts the opportunity to publish their projects openly at the UTA MavMatrix, which includes any data collected and software/analysis conducted. The DFRL and the  Research Data Portal at UTA provide guidance and education on data management throughout the research lifecycle.

You can also make an appointment with the data experts at UTA Libraries:

• John Connolly – Data Scientist/ Statistical Consultant
• Ryan Hedrick – Data Management Librarian
• Rubab Shahzad – Data Visualization Librarian

Follow our Open@UTA Blog series and learn more about Open at UTA.