,
27 tweets,
8 min read
Read on Twitter
I'm going to live stream the creation of a scientific paper, including both #Rstats coding and writing.
Here's why and how...
[THREAD]
Here's why and how...
[THREAD]
Watching other people play computer games has become phenomenally popular. If you think this sounds absurd then you're probably over the age of thirty or don't have any kids yourself. E-sports viewer numbers will soon overtake conventional sports medium.com/@IGGalaxy/espo…
Along with e-sports, there’s a smaller community of people that live stream their coding and writing. I've tried live streaming a few R coding sessions myself. These were daunting (but fun) experiences
But these live streaming sessions felt a little artificial, as I was demonstrating coding tasks I'm fairly familiar with.
This isn't how science works. Datasets are messy, you're learning new skills on the fly, and wrestling with cryptic error codes xkcd.com/979/
This isn't how science works. Datasets are messy, you're learning new skills on the fly, and wrestling with cryptic error codes xkcd.com/979/
You only tend to hear about people successes in academia. But considering that success rates for most grants and paper acceptances in prestigious journals are hovering around 10-20%, *someone* out there must getting rejections—we're just not hearing about it.
Outright successes & failures are two ends of a spectrum, which occasionally punctuate our careers. The reality is that a typical day in academia is quite banal. You don't often see tweets about these run-of-the-mill of days, despite their ubiquity
Now that the technology is available and that there seems to be some interest in watching other people work, I want to demonstrate what writing a paper is *really* like (the good, the bad, and the banal) by live streaming the creation of one.
There are three reasons why...
There are three reasons why...
1. To offer a peek behind the curtain
There’s a lot of mystery surrounding how the writing process. Young scholars are just *expected* to learn how to pick up this difficult skill without any formal teaching
There’s a lot of mystery surrounding how the writing process. Young scholars are just *expected* to learn how to pick up this difficult skill without any formal teaching
Of course, I don’t think the way that I write papers is the best way, but it’s certainly *one* way.
I enjoy making tutorials about R and scientific writing and think this is an important service to the community as an early-to-mid career scholar, but these things take time.
I enjoy making tutorials about R and scientific writing and think this is an important service to the community as an early-to-mid career scholar, but these things take time.
By live streaming the work that I’m *already* doing, I'll be making better use of my time, which is especially important to me now that I'm a parent, which introduces its own time-management challenges
2. Accountability
I want to see if live streaming will help me work more efficiently knowing that people *can* be watching. I already block access to social media, but what’s harder to monitor is ‘false hustle’ work, like updating my CV.
I want to see if live streaming will help me work more efficiently knowing that people *can* be watching. I already block access to social media, but what’s harder to monitor is ‘false hustle’ work, like updating my CV.
I have no idea if anyone's going to watch this stream, and I frankly don’t really care. I'm more interested to see how this extra level of accountability will influence my work.
3. To improve how I write manuscripts and code
I want to get better at writing and definitely get better at coding. When it comes to coding, I've got a grasp of the basics but I know my code is inefficient and often has redundancies.
I want to get better at writing and definitely get better at coding. When it comes to coding, I've got a grasp of the basics but I know my code is inefficient and often has redundancies.
I gotten some great tips from people who've watched my live streamed R coding sessions that save me time
When it comes to my writing, I'm interested to see whether I can get some real-time (or close to real-time) comments as I write my manuscript. Co-authors can offer some great feedback, but they often share our the biases that are easy to miss.
My paper topic:
I think the current push for open data is the right way forward, but there are many circumstances where it's not possible to share your data due to privacy concerns. Our research participants are entitled to their privacy
I think the current push for open data is the right way forward, but there are many circumstances where it's not possible to share your data due to privacy concerns. Our research participants are entitled to their privacy
There is a clear trade off to be made between the utility of sharing data and disclosure risks. Of course, if we were to share ALL of our data this would be incredibly useful for science, but blanket data sharing would be at the cost of disclosing personal information.
The 'synthpop' R package (what a cool name!) seems to offer an ideal solution for increasing the utility of shared data without sacrificing disclosure risk. It creates a synthetic version of your dataset while retaining the same properties and relationships between variables.
There are a few reasons why I'm really interested in synthetic datasets...
1. Readers and reviewers can better understand the data, as they can explore dataset properties such as distributions, variance, and outliers.
1. Readers and reviewers can better understand the data, as they can explore dataset properties such as distributions, variance, and outliers.
2. Other researchers can explore the data and fit hypothesis-generating models, which can be verified by the original authors using the real dataset.
3. Synthetic datasets can be used to accompany analysis scripts, which can be difficult to follow without the dataset
3. Synthetic datasets can be used to accompany analysis scripts, which can be difficult to follow without the dataset
Culture and behavior change is hard. An important step for the wholesale adoption open science practices is to MAKE IT EASY cos.io/blog/strategy-…
My paper plan is to demonstrate how to create synthetic datasets and validate them. Synthpop has been used to create synthetic datasets in the biobehavioral sciences once or twice before, but right now there's no accessible guide for using synthpop in this context
I will focus on oxytocin research as that's my field of study and that datasets from oxytocin studies are rarely shared. I'm planning on using 2-3 datasets that are already open. I might also introduce missing data, outliers, and skewness to examine the robustness of the methods
I plan to live stream twice a day (mornings/afternoons) during the weekdays beginning July 22, in 45 min sessions. I might do more or I might do less, depending on how my day is going. I'll use both Twitter and Twitch for streaming and then post the video afterwards on my YouTube
Here's the (slightly longer) blogpost version of this thread dsquintana.blog/why-livestream/
And yep, you can count watching this as work 😉
🔴 Going live in 30 minutes, let's see how this goes...
