feedback on NSF grant budget

[tdhock]

Hi all! This issue is about creating trust/transparency between the NSF grant team, and the larger data.table community. I am the Principal Investigator (PI) of the NSF grant, which means I am in charge of managing the money, and making sure it gets spent according to what we outlined in the budget. The goal of the grant is to expand the open-source ecosystem of users and contributors around data.table, so that hopefully by the end of the grant, the development of the project can continue and thrive without any more money from the NSF. I have already written a blog post about the main grant activities https://rdatatable-community.github.io/The-Raft/posts/2023-10-15-intro_to_grant-toby_hocking/ Here I would like to further explain the grant budget, and invite feedback about how we should spend the money. (within the constraints imposed by the budget) First of all, final budget and justification which was approved by NSF is: final-budget.pdf It is for US$731K from summer 2023 to summer 2025, to be spent as follows:

• US$242K indirect costs and US$79K fringe benefits (go to the university, we do not get to spend it directly in this project)
• US$78K pays summer salary for myself and the two co-PIs of the project, @igorsteinmacher and @marcogerosa There is some room for discussion about how we spend this time. My main activities are supervising the overall project, writing the governance, and implementing the revdep check system, https://github.com/Rdatatable/data.table/wiki/Release-management-and-revdep-checks . Igor was ~leading~ active in the NSF interviews #5880 and I expect that Igor and Marco will lead activities related to evaluating the effectiveness of the NSF project, near the end (writing papers about what techniques were effective for increasing contributions, etc)
• US$123K pays a research software engineer @Anirban166 full time for 1 year and a master student @DorisAmoakohene part time for 2 years. They are supposed to be working on creating tools that will help make it easier for new users/contributors, and so far the main focus has been performance testing https://github.com/rdatatable-community/The-Raft/pull/16 and Ani has been leading the interviews https://github.com/rdatatable-community/The-Raft/pull/21 but we would welcome feedback about how the community thinks these two could more effectively spend their time, to expand the data.table open-source ecosystem.
• US$96K pays Kelly Bodwin @kbodwin and Tyson Barrett @TysonStanley to work part time as consultants over project. Kelly has been working on a blog https://rdatatable-community.github.io/The-Raft, social media, and a new bookdown data.table textbook. Tyson has been working on interviews #5880 as well as having taken over CRAN maintainer role. Again we would welcome feedback about how the community thinks these two could more effectively spend their time, to expand the data.table open-source ecosystem.
• US$10K pays amazon web services for compute time, to test performance benchmarks similar to https://duckdblabs.github.io/db-benchmark/
• US$10K pays translators, none funded yet but you can apply via: https://rdatatable-community.github.io/The-Raft/posts/2023-10-17-translation_announcement-toby_hocking/
• US$70K pays travel expenses for NSF project personnel (PIs and consultants). So far I met with Kelly and Tyson at JSM'23 in Toronto, and I gave a tutorial at the LatinR conf in Montevideo, https://rdatatable-community.github.io/The-Raft/posts/2023-11-19-LatinR_summary-toby_hocking/
• US$21K pays travel awards for non-NSF project people https://rdatatable-community.github.io/The-Raft/posts/2023-11-01-travel_grant_announcement-community_team/ so far we have approved one person, @jangorecki who recently gave talks in Edinburgh and Sevilla https://github.com/Rdatatable/data.table/wiki/Presentations

[waynelapierre]

Could you provide more details on this ? US$242K indirect costs and US$79K fringe benefits (go to the university, we do not get to spend it directly in this project).

[tdhock]

hi Wayne thanks for asking. It is normal to ask about indirect costs, if you are not familiar with the university/federal funding system. This page https://in.nau.edu/osp/frequently-requested-information/ gives several FAQ answers, including "2. How do universities spend indirect cost payments from the federal government? Indirect cost payments are actually reimbursements for costs that universities have already paid for expenses incurred in conducting federally-sponsored research..." The fringe benefits are explained in the budget justification, "Employee-related expenses (ERE) are estimates based on the projected cost of health, dental, life, disability, FICA and Medicare, unemployment..."

[Anirban166]

@tdhock I was the one leading the NSF interviews so please update the misleading statements in your post above.

To be more precise, I led and did the majority of them (50+ out of the minimum required count of 60, reaching out to several more), plus was actively involved in whatever else NSF wanted the data.table team to do for the duration of the program. That includes making and giving multiple comprehensive presentations representing different aspects of data.table (hypotheses regarding contributors' motivations and the like, governance, sustainability model, etc.), along with appropriate diagrams (such as an ecosystem map and BMCs).

[tdhock]

@Anirban166 has started a blog where he plans to post every day about what he is working on to help expand the open-source ecosystem of users/contributors around data.table, https://anirban166.github.io/data%20table/ as well as information about how you can help, and give feedback.