Heidi Seibold - Are (data) scientists bad at science?

Let's face the facts: most research is not reproducible. That means, if you run the same analysis on the same data of a (data) science project again, you will likely not receive the same results. Ups! In this talk I want to explore the stumbling blocks that (data) scientist run into and what to do about it. Spoiler: it has a lot to do with getting organized.

Transcript

1. CC-BY 4.0 Are (data) scientists bad at science? Munich Datageeks May Edition 2023 Heidi Seibold Heidiseibold.com Slides: https://bit.ly/3IvAzT0
2. Please ask questions!
3. Image: https://the-turing-way.netlify.app
4. Is this what we want? DOI: 10.1371/journal.pmed.0020124
6. https://the-turing-way.netlify.app/reproducible-research/overview/overview-definitions.html
10. How can we do better?
11. What are you doing well already?
13. Project management = good organisation Let's not pretend: we're not geniuses ;P http://www.quickmeme.com/meme/3r98zx
14. Good organisation … starts simple • Good naming • Nice file organisation
15. Naming • Myabstract.docx • Joe’s Filenames Use Spaces and Punctuation.xlsx • figure 1.png • fig 2.png • JW7d^(2sl@deletethisandyourcareerisoverWx2*.txt • 2014-06-08_abstract-for-sla.docx • Joes-filenames-are-getting-better.xlsx • Fig01_scatterplot-talk-length-vs-interest.png • Fig02_histogram-talk-attendance.png • 1986-01-28_raw-data-from-challenger-o-rings.txt NO YES See slides by Jenny Brian
16. Naming • 2014-06-08_abstract-for-sla.docx • Joes-filenames-are-getting-better.xlsx • Fig01_scatterplot-talk-length-vs-interest.png • Fig02_histogram-talk-attendance.png • 1986-01-28_raw-data-from-challenger-o-rings.txt YES File names should be: ➔ Machine readable ➔ Human readable ➔ Optional: Consistent ➔ Optional: Play well with default ordering
17. Organise your files and folders well . ├── analysis <- all things data analysis │ └── src <- functions and other source files ├── comm │ ├── internal-comm <- internal communication such as meeting notes │ └── journal-comm <- communication with the journal, e.g. peer review ├── data │ ├── data_clean <- clean version of the data │ └── data_raw <- raw data (don't touch) ├── dissemination │ ├── manuscripts │ ├── posters │ └── presentations ├── documentation <- documentation, e.g. data management plan └── misc <- miscellaneous files that don't fit elsewhere https://github.com/HeidiSeibold/research-project-template
18. What can data scientists do?
19. What can data scientists do? • Work according to good practices • Be a role model • Collaborate • Teach
20. heidiseibold.ck.page/posts/f eedback-wanted-building- a-digital-research-academy
21. More… Slides: bit.ly/3IvAzT0