class: title-slide center middle inverse # Agile Research ## How to work efficiently, even from <i class="fas fa-home "></i> <br> ### Florian Privé #### NCRR WIP -- October 2, 2020 --- ## About me <br> My background to understand where this talk is coming from: -- - Engineer in Informatics and Applied Mathematics from France `\(\rightarrow\)` I could have been developing software for companies instead of doing Research -- - PhD in predictive human genetics -- - Just started my second year as a postdoc at NCRR -- - Now have the perspective as both first author and (co-)supervisor --- ## Agile workflow <br> - Primarily developed for software development - A response to the problem of many teams failing to deliver a working product in time, and funders pulling money out because they lost confidence or/and interest in the project over time <br> -- `\(\Rightarrow\)` What is the best workflow to turn an idea into reality? --- ## Agile workflow in Research? <br> A crude analogy: -- - the team `\(\longrightarrow\)` the first author(s) -- - the final product `\(\longrightarrow\)` the paper -- - a working component of the product `\(\longrightarrow\)` a figure or table of the paper -- - stakeholders/funders `\(\longrightarrow\)` the supervisor(s) -- <br> `\(\Rightarrow\)` What is the best workflow to turn an idea into a paper? --- ## The main component of an Agile workflow: the sprint <br> -- - One element of the *TODO* list, which **should not take very long** (say one week or two max). -- - You should try to further subdivide this sprint into even smaller tasks. -- - Once you start one of these tasks, you should try to **focus** only on this, until completion. -- - It is important to always set a **maximum time limit** for a task or a sprint. -- - If you can't finish in time, maybe you need help? -- - If you are exploring an idea, maybe it just does not work and you should move on to another idea? -- `\(\Rightarrow\)` try to **fail fast** and move on <i class="fas fa-info-circle "></i> only 1 out of 10 of my promising ideas actually end up in a paper! --- ## Another Agile principle: deliver fast and regularly <br> I'm not saying you should draft a paper within two weeks but.. <br> -- - Deliver a proof of concept quickly, or move on to another idea -- - Deliver regularly, e.g. one figure or table that could go into the paper -- - Keep the supervisor(s) enthusiastic and confident about the work, and keep you motivated as well -- - Get feedback sooner to avoid going in the wrong direction for weeks -- - Welcome updates on your analysis `\(\Rightarrow\)` this will make your code more reproducible if you know you may have to update it later. -- <br> `\(\Rightarrow\)` Agile provides speed and energy! --- ## Also part of the Agile workflow: the stand-ups <br> -- Everyday, everyone quickly says - what they have been working on - what they will be working on (e.g. planning the task(s) of the day) - if they are stuck somehow (need help, feedback maybe?) -- <br> - Bjarni, Clara, Emil and I had these stand-ups on Zoom during the first lockdown. - It helped keeping everyone on track while working from home. - Should continue doing it? Not always possible? Maybe opt for a written alternative? --- ## Summary and conclusion <br> -- - Design small tasks that you can do quickly. -- - For exploratory analyses, set a time limit beforehand. -- - Deliver as early and regularly as possible, to keep everyone on track and motivated by the Research project. <br> -- - The Agile methodology is a tried-and-tested method in the Industry. -- - This is basically how I try to work and how (I think) I remain effective, even when working from home. <br> -- <i class="fas fa-info-circle "></i> I am moving back to France in 2 months, and I'll be working remotely from there for the last 10 months of my contract. --- class: center, middle, inverse # Thanks <br> Presentation available at https://privefl.github.io/thesis-docs/agile-research.html <br> Some inspiration from *Agile Data Science with R: a workflow* by Edwin Thoen, available at https://edwinth.github.io/ADSwR/ <br> <br> <i class="fab fa-twitter "></i> [privefl](https://twitter.com/privefl) &nbsp;&nbsp;&nbsp;&nbsp; <i class="fab fa-github "></i> [privefl](https://github.com/privefl) &nbsp;&nbsp;&nbsp;&nbsp; <i class="fab fa-stack-overflow "></i> [F. Privé](https://stackoverflow.com/users/6103040/f-priv%c3%a9) .footnote[Slides created via R package [**xaringan**](https://github.com/yihui/xaringan)]