- 4
- 31 019
Selina Baldauf
เข้าร่วมเมื่อ 26 ม.ค. 2020
AI tools for scientific coding and data analysis (Github Copilot + R)
AI tools promise to enhance our programming workflows by making them faster and more efficient. When used correctly, these tools can assist us in writing higher-quality code, increasing productivity, and acquiring new techniques easily. In this lecture, I will briefly discuss AI tools for programming in general, followed by an introduction to the integrated AI tool, Github Copilot. I will demonstrate how this tool can help you write code more quickly, identify and correct errors in your code, generate automated documentation, and provide assistance with good coding practices.
Additionally, I will provide a brief demonstration of Codium AI. During the demonstration, I will showcase a data analysis project in R, utilizing VS Code as the IDE. However, it is important to note that these tools can also be used with other programming languages and IDEs. Finally, I will briefly touch on some guidelines and concerns related to utilizing such tools in scientific contexts.
A Step-by-step guide on how to get Copilot for free for Visual Studio Code or RStudio: selinazitrone.github.io/tools_and_tips/sessions/additional_material/07_ai_tools/get_copilot_step_by_step.html
GitHub Copilot Docs: docs.github.com/en/copilot- Privacy FAQ of
GitHub Copilot: github.com/features/copilot/#faq
Prompt engineering with GH Copilot to get better results: dev.to/github/a-beginners-guide-to-prompt-engineering-with-github-copilot-3ibp
Codium AI: codium.ai/
Rtutor - Browser-based tool where you can ask questions about R and data analysis. You can also upload a dataset and ask the AI to analyse it. You can then download the corresponding R script or Rmardown report: rtutor.ai/
Nature living guidelines on responsible use of generative AI in research: www.nature.com/articles/d41586-023-03266-1
Additionally, I will provide a brief demonstration of Codium AI. During the demonstration, I will showcase a data analysis project in R, utilizing VS Code as the IDE. However, it is important to note that these tools can also be used with other programming languages and IDEs. Finally, I will briefly touch on some guidelines and concerns related to utilizing such tools in scientific contexts.
A Step-by-step guide on how to get Copilot for free for Visual Studio Code or RStudio: selinazitrone.github.io/tools_and_tips/sessions/additional_material/07_ai_tools/get_copilot_step_by_step.html
GitHub Copilot Docs: docs.github.com/en/copilot- Privacy FAQ of
GitHub Copilot: github.com/features/copilot/#faq
Prompt engineering with GH Copilot to get better results: dev.to/github/a-beginners-guide-to-prompt-engineering-with-github-copilot-3ibp
Codium AI: codium.ai/
Rtutor - Browser-based tool where you can ask questions about R and data analysis. You can also upload a dataset and ask the AI to analyse it. You can then download the corresponding R script or Rmardown report: rtutor.ai/
Nature living guidelines on responsible use of generative AI in research: www.nature.com/articles/d41586-023-03266-1
มุมมอง: 1 779
วีดีโอ
Effective digital notetaking with Obsidian
มุมมอง 22Kปีที่แล้ว
Effective note taking is crucial for both students and scientist. It’s useful to manage tasks, capture key insights from papers, and write down new ideas. In this lecture, I will introduce you to my note taking software of choice. Obsidian is a highly customizable, markdown-based tool that lets you organize your notes efficiently and create interconnected webs of thoughts and ideas (also called...
Efficient R
มุมมอง 6Kปีที่แล้ว
For most data analysis and statistical computing, R is efficient enough. However, there are times when we encounter bottlenecks in our code that slow it down significantly. In this lecture, I’ll teach you techniques to identify those bottlenecks and write more efficient code. You’ll learn the fundamental principles of faster R code and discover efficient packages for data analysis. We’ll also t...
Game of Life with C++ and Qt
มุมมอง 1.1K4 ปีที่แล้ว
You can find the GitHub repository here: github.com/selinaZitrone/GameOfLife_Cpp
I appreciate your honest unglazed representation.
Thank you. So informative !!
Thanks for this lecture. I start using the collapse package
Thank you so much for such an informative video.
Efficient R is an oxymoron surely
@@generalyoutubewatching5286 Yes this argument can be made for sure 😃. Maybe it's more about "efficient enough R" or "more efficient R". But for sure R should not be used if efficiency is the main requirement for a project 😉 But at least in my field (biological science) it's efficient enough for most tasks
Excellent tutorial and to the point. Please keep up posting!
👍
Amazing video! I work as an econometrician and your presentation gives me a lot of ideas 😊 Could you share your presentation as a pdf file?
Thanks for your nice feedback, good to hear that you got ideas from it! A bit clunky, but you can download the presentation pdf from the depths of my Github project: github.com/selinaZitrone/tools_and_tips/blob/main/docs/slides/2023_11_16_efficient_r.pdf
🙏
Thank you for the simple and straightforward video. Obsidian is an amazing app and I have so much more to learn about it
Very nice!
Great Video
Thank you for your effort ❤
Very interesting video! I am using Obsidian in my TH-cam channel for organizing my notes and it's very cool :D
Great video. I didn't know a lot of the packages, thanks for the clear explanations.
Is there an online formal course with Selina Baldauf?
@@fitzg3278 No there is no formal online course for this topic. I give some workshops at my university and at conferences but not in general as a formal online course.
How great is that!! I never thought that TH-cam would one day bring me to see the answers to the questions I was looking for from an old colleague. Many thanks, Dr-Selina
Special thanks from a brazilian economist! Learned a lot.
Great and useful video, thank you so much!
Good video! In the matter of practices for effective research. Do you have more recommendations from your own experience, books, articles, other YT channels on how to revise papers, more ideas on making the annotations, and develop research ideas?
Good question. There is a lot of material out there and I feel like a good workflow is is very individual. My main recommendation is: Don't let all the complex workflow videos/articles stop you from just starting to take notes. Start with a simple workflow and don't be afraid to change and adapt it. In the beginning, I tried to build the perfect workflow from scratch which was impossible and stopped me from actually taking useful notes. But some of the resources that I liked and that helped me get a clearer idea for myself are: - Book "How to take smart notes" by Sönke Arendt - This blog post: ilyashabanov.substack.com/p/note-taking-system-for-success-in - YT Videos on Zettelkasten by Artem Kirsanov, e.g. this one: th-cam.com/video/E6ySG7xYgjY/w-d-xo.htmlsi=1xd753lz0GcPOO-w But other people feel free to also post your favourite resources as an answer here :)
Awesome tutorial, thanks for this! :)
Very good tutorial. Thanks!
Amazing video. Thanks you, I will start learning from u.
I've been looking at many videos but leaving me only frustrated. But you actually showed how it's done, the best video I've seen on this yet!
Hi, Really appreciate the effort here, an interesting workflow i would say. Some questions regarding zettle kasten even after goinf through notebook_demo: 1. what would be the permanent notes in this workflow? a standalone zettle with some reference to researtch paper? 2. if a new idea originates after taking literature notes (reading multiple research papers), should we make a zettle note for that idea and link all research papers or literature notes to that? 3.based on above question how would be manage projects with ideas? e.g we might read alot of papers for specific project now to make workpackages we can use tasks but to work on some new approaches should we define it in form of zettle ? Thanks, would appreaciate a response
Hi, thanks a lot for your feedback. Very good questions you have there :) I would first of all say that there is no single right way to handle this. Of course there are suggestions for workflows but ultimately, you have to find something that makes sense to you. How I would answer your 3 questions from my workflow: 1. A permanent note for me is a Zettel about one concept/idea/theory/... If the concept is bigger, I split it into separate notes and then have one meta-Zettel combining them together. And in the Zettel, I just put links to my literature notes that I used. In my workflow literature notes can also be about YT-videos, blogs etc. Not necessarily only research papers. 2. Yes, I do it like that. I create a new Zettel and link the research papers. 3. If I collect ideas for a project, I always have a project file that links to individual notes with Ideas for that project. Some ideas go into Zettel (mainly the ones I think are general/relevant for other projects as well), but some I also keep on separate project notes that are also linked in the main project file. This is mainly for ideas/thoughts that I only want to have in the project but not keep forever in my Zettelvault. That said, I can only encourage you to start with the Zettel system of notetaking and adapt it to your own preferences over time. It needs some time to find a workflow that feels good and that is useful for you. If you are interested, I can also recommend this video that gave me a lot of inspiration: th-cam.com/video/E6ySG7xYgjY/w-d-xo.html th-cam.com/video/D9ivU_IKO6M/w-d-xo.html Hope that helps :)
@selinabaldauf7529 thanks for taking your time and answer my queries. Especially the advice in the last part. I think I will give it a go and see. Nevertheless superhelpful content 👌
How did you coloured your sidebar?
I use the AnuPpuccin theme that you can install via the Obsidian settings. Then I use a community plugin called "Style settings". In this plugin you can change a lot of settings of the AnuPpuccin theme, including the coloring of the folders. I saw this in a TH-cam video by Nicole van der Hoeven: th-cam.com/video/7-SOwxpZQNI/w-d-xo.htmlsi=676FpBIDsPwWYJIe
@@selinabaldauf7529 oh thx
Great video, Selina, thank you for sharing.
Finally a channel that shows how to improve your scientific work and organization! Great video!
what videos are best for someone new to r and programming in general.
Good question. I would say it depends on what you want to learn. E.g. do you want to use R for programming or for data analysis? I don't know the best videos for learning R, but two (free) books I recommend to my students to get started with are and data analysis are: 1) Hands-On Programming with R - Gives a nice intro to R from the very beginning: rstudio-education.github.io/hopr/ 2) R for Data Science - Learn how to use R and the tidyverse packages for intuitive data analysis (from reading data, cleaning, transforming it to visualization): r4ds.hadley.nz/intro If anyone knows good videos, feel free to post them below :)
@@selinabaldauf7529 Thank you! 😊For data analysis and because there is so many people using data for propaganda purposes it would be nice to be able to check and analyse the info myself, where should I start? maths? statistics? or straight to R, I'm not sure but how would you fact check dubious data claims that get shared on social media
This is a very big question that I think does not have a single good answer. If your goal is only debunking, then I think the best and fastest option is to look for reliable fact checker websites/accounts that already do this with a lot of experience. Debunking the information yourself requires you to get the raw data (which can already be tricky sometimes), know the right statistics and know an analysis software like R. So that is quite an investment of time. But of course also fun to learn, so don't let me stop you 😊
wow this was very clearly explained, keep doing tutorials because you explain things in a way anyone can understand. 😊
Thanks a lot. Clear and understandable for everyone was the goal 😊
Obsidian is a proprietary software. Don't promote it. Better use Logseq or Joplin (or other - there are plenty of open source alternatives).
That means do not write in R.
Well, I guess it depends on what you want to do 😊 but if speed is the priority, R might not be the right choice
Is it mandatory to use R within VS?
No it's not. You can use many different combinations of programming languages and IDEs. If you are e.g. an R user and you use RStudio, you can also integrate Github Copilot there (although without the chat functionality). Check out the step-by-step guide that I linked in the video description. There I describe how to get Github Copilot and I link how you can set it up with VS Code, but also with R Studio.
Best tutorial ever. Thank you very much 🙏🏻🌹
Very informative on how to optimize R code
Wow, this was a great introduction to Obsidian. Very impressed by the clarity of diction & very pleasant & engaging voice. (& incredible high rez audio) The phrase <walkthrough of the workflows>, some serious caffeine inspiration right there!
You have new sub
appreciate your efficient explanation of your system! I especially found clarity around how you described your information management systems; project, personal, literature, etc. This helped me better organize my own system, much more to my thinking than simply PARA.
Thanks for your feedback. Great to hear that it helped you 😊 It took me some time to figure out this organization system but now that it's set up, it saves me just so much time and frustration
THANK YOU - one of the best videos I have had on R in some time
oh hey i loved the complete comprehensive guide and everything amazing fidelity & intuitive teaching! a recommodation - pdf++ plugin/extension is great to annonate check it out which would true gamechanger - it can help you annotate with colors and it would link to the note and it also has selection or like snapping any images or anything
Thank you for the video. I really recommend using Obsidian + Copilot plugin + Ollama to bring note taking to the next level!
This is brilliant. I've been using obsidian a bit, but I knew that I was definitely not making the most of it. This presentation has many great ideas in it and I'm looking forward to putting them into practice! Thanks much!
Thanks from Ghana.
Thank you so much, it is clear, to the point and easy to follow. You are a great teacher!
Thanks a lot for your nice feedback. Happy note taking!
Thank you. That’s a very nice tutorial!
I'm glad that youtube recommended your channel, first the efficient R video and now this software, thanks for sharing your knowledge with us!
Thanks for your nice feedback. Happy you liked the videos
This was great! Thanks Selina!!
*sf::st_intersection(),* the epitome of least efficient code in the universe!
Do you by any chance know a more efficient function that does the same thing?
@baldauf7529not in R. But performing the same operation in GUI-based packages like QGIS is lightning fast compared to this function. I haven't managed to know why this is so slow at this level.
By far, my efficiency hurdles with R are in terms of memory usage. I don't really understand why R needs to load all the dataset into memory before working with it. I am sure the clever guys at Posit could address this issue if they chose to.
I didn't know about the collapse package, this is an amazing discovery ! Thank you for this content