Data Science Hangout | Travis Gerke, PCCTC | Tips for *Very* Remote Work

Hi everybody, welcome to the Data Science Hangout. I'm sure we'll have quite a few people joining in in the next few minutes here too. But if you're joining for the first time, it's great to meet you. I'm Rachel. If this is your first Hangout, this is an open space for the whole data science community to connect and chat about data science leadership, questions you're facing, and what's really going on in the world of data science. These sessions are recorded and shared to YouTube as well as the RStudio Data Science Hangout site. So you can always go back and rewatch to find helpful resources or share with others.

I'll say this up front because I always forget this. We also have a LinkedIn group for the Hangout too. So if you ever want to continue a certain discussion or see a brief summary of a topic brought up each week, you can join that group. I'll share it in just a second so I don't pause here. But we always want to create spaces where everybody can participate and we can hear from everyone. So there's three ways you can ask questions today. You can jump in by raising your hand on Zoom and I'll just call on you. You can put questions in the Zoom chat and feel free to just put a little star next to it if you want me to read it out loud instead. Or I can call on you to introduce yourself and ask. But we also have a Slido link that Hannah will share in just a second here in the chat where you can ask questions anonymously too. So this is all for you and all the questions are audience led. So just like to reiterate, we love to hear from everyone, no matter your level of experience or area of work.

But today, I'm so happy to be joined by my co-host, Travis Gerk. Travis is Director of Data Science at the Prostate Cancer Clinical Trials Consortium. That's a hard one to say. But Travis, I'd love to turn it over to you and have you just introduce yourself and maybe share a little bit about the work that you do.

Yeah, sure. So hi, I'm Travis. Thanks so much for having me. This is really great. I've been on a couple of Hangouts and they're always fun. So right, as Rachel said, I'm Director of Data Science at the Prostate Cancer Clinical Trials Consortium, which yes, it is a mouthful. An acronym might be even more tricky, PCCTC. Try saying that five times fast. But whatever the case of you, we're sort of a we're a CRO, so a clinical research organization, meaning that we facilitate the work of clinical trials from end to end for often pharmaceutical sponsors or sometimes so-called independent investigator initiated trials, which will be maybe physician scientists who've initiated a trial at their at their site at their hospital. And we have a we're sort of a boutique CRO in the sense we mostly do prostate cancer oncology work. So right now we're running about a quarter of all early phase clinical trials in late stage prostate cancer in the US. And we have a few global studies as well.

So it's good. We do, as I said, we kind of collect all the data at the sites. We help the sites launch the trial. And then once the data starts rolling in, that's when the data science team takes over. They do a lot of counting, a lot of tabling, big figures. We use R Markdown. Something that we're excited to do is move the clinical trials reporting space more towards R, Python, other open source technologies over traditionally a licensed dominated space, such as SAS and things like that. Of course, we're not, we're not the only kind of thought leaders in the space. There's a lot of great work from GSK, Roche and Taurus and others that we keep tabs on. And we love to learn from you all as we go.

Very remote work: life in an RV

Thank you. And so I know a lot of people have commented in the chat already about your amazing background and wanting to know, where are you right now?

I've seen the guesses you're coming, Glacier National Park. Like I have, I feel like I've become Bigfoot these days. There was someone on Twitter recently said, did I just have a Travis Gerke sighting? This, no, I'm not in Glacier anymore. We're now in Washington State. I'm on a river called the Nisqually, which is just outside of Olympia, near Seattle. Awesome. So I guess I'm going to transition. No, I was going to say, I feel like that's a great transition into, I hinted that we would chat a little bit about very remote work and would love to learn a little bit more about your travels and working remotely. Yeah, this is something kind of wild that we did. So my wife and we have two kids, they're seven and eight and a dog.

We used to live in Florida and then we were never fans of the heat. And then for that and reasons, we just said, let's leave. So so we sold our home and we bought an RV and a truck and sold most of our stuff. And now we live full time in an RV. And I just do data science from from wherever we are. Lately, it's been on the Pacific Northwest, but it's great. I'm happy to answer any odd questions you might have about that. I 100 out of 10 recommend it if anyone has the urge to just get up and move. It's very pandemic friendly. We don't see many people often, although we want to go to the city. We can. We went to Seattle yesterday. It was fun. So yeah, it's one of the good things about data science. It's pretty flexible. You can do from anywhere, really.

I 100 out of 10 recommend it if anyone has the urge to just get up and move.

Staying connected: internet on the road

That's amazing. I see, Libby, you asked a question somewhere a bit earlier about Internet. Do you want to jump in?

Yeah. Hey, Travis and I are talking about this on LinkedIn a little bit. I was like, you're going to spill all your secrets about how you have Internet and all these places, right, because I my husband and I both work remotely and we both feel like the Internet is like a utility for us now, like a boundary to to working. So I'd love to know more about like how you stay connected all over the place. And if you have any trouble.

Yeah, no, I've I've shockingly actually not had any trouble yet. I geeked out on this big time before we did this, because you're right. Without Internet, can't do much in terms of work. But even in the remotest places we've been, I think Big Bend in Texas, we were the furthest from civilization. I was still able to maintain Zoom meetings, much to my surprise. Actually, I prepped for a little bit on a week like that where I'm pretty remote. I make sure to set up a lot of kind of maybe it's a learning week or, you know, kind of heads down work. It's actually kind of good. It gives you a little bit of space from the usual grind of just Zoom to Zoom. But even in that case, it wasn't necessary.

So today I'm using this hotspot. So these hotspots, this one's the 5G Netgear. I can always share details about these things. But this one. So I have redundant phone plans for this sort of thing. This one today, I'm on AT&T. This hotspot also runs a T-Mobile. But I also have Verizon plans for it just depends. It depends where we are, what kind of service is going to give us the best. But I've actually never had a challenge streaming videos, Zoom, things like that. Something else that's helpful if you get a hotspot, these little boosters. This is like $20 on Amazon. They can be tremendously helpful. They kind of like funnel the signal into your hotspot a little bit better. There is one. And if you want to really geek out on this, like I did, the best resources website called RVMobileInternet.com and you can check that out. These days, I think if I even had like a proper home, I might consider doing this instead of proper like cable or whatever internet, because the price is pretty similar. And often I have better signal much if I'm using it than my kind of at home colleagues that I'm Zooming with. So it works well.

Working outdoors and managing time zones

Aaron, I see you asked a question, too. Do you want to jump in?

Hi, this is Aaron. Um, I was just wondering, because I've tried to work in places where there's lots of sunlight before, and I found it's really hard to, even if the sun isn't shining directly on you, if there is any sun touching anything in the background behind you, it's hard to keep the glare off the screen. So I was wondering if you just had tips for us.

I wish I had a brilliant thing to suggest and maybe others will. For today, I kind of just wandered around until I found a shady spot. So I'm kind of like in the shade and I'm actually letting just my MacBook Pro do all the work. I'm even using the onboard camera. I mean, I do have, I do have like a stream cam that sometimes I use, but most times I find this is okay. I'm sorry if that's a disappointing answer.

I see, Mike, that you asked a question in the chat. You want to jump in?

Hi, thanks. Thanks for discussing this, Travis. I hope you don't mind that we're talking about, you know, your, your work lifestyle rather than data science or anything complicated. So I, I work remote from the most of my team. They're predominantly US based and classic situation here in the, in the UK. Today is a bank, a public holiday because of the Queen's Platinum Jubilee. Last Monday was a US public holiday. So, you know, my colleagues are hitting me up on my public holiday, but I wasn't allowed to hit them up on their public holiday, but when you're out and about, how do your colleagues kind of, how do you, how do you balance up, you know, time at work versus, you know, time I can be available for conversations and how do you have those conversations with colleagues to say, look, you know, I'm going to be busy or traveling between here and here, you know, so don't expect an answer in that time period.

Yeah, it's such a great question. And it is always, I mean, it's something that one has to be very intentional about and just communicate broadly. I do work very odd hours, especially now that I'm on West coast. Most of my colleagues are on East coast time. So I do try mostly to wake up very early and stick to their schedule, but a lot of times it doesn't work out. And that's fine. I'm kind of a night owl by nature anyway. So I will fire off emails at what is going to be like 2am Eastern. And what I do make sure to communicate is that, you know, if I'm, if I'm sending an email at 2am, like it doesn't mean I expect anyone to see or respond or do anything with that email at 2am. Nor does it mean that I've been working from like 9am Eastern to 2am Eastern. Like, it's just, you know, the, the async, I think, I mean, that's really the future of most work environments for data science, I would suppose. I think someone like Chris Hurd on Twitter probably has stronger words than I do about that, but which, which I would recommend as a follow for remote work hot takes, but yeah, I mean, it's just, it's, it's really comes down to communication. I let people know I'm going to be traveling and, you know, blocking time on my calendar, things like that. Cause some afternoons I do, we do need to move or make it, make a drive or something like that. And, you know, I just, I just set it as busy and then people know that I'm, I'm not there and then just let them know.

I think, I think blocking calendars is the, the, the key, right. Cause that says, you know, don't expect an answer in this three period, three hour period or whatever. I like that point though, about you might answer it too, but that doesn't mean you're expecting somebody to answer you right away too. It's when people answer back and you get into a conversation. It's Katie from our team used to put something in her email signature. Like I send emails at off times, but that doesn't mean I expect a response.

Yeah, that was, we use that tagline a lot of my last job. In fact, our president had it on his email. I'm working on a remote schedule. I know I don't expect an answer back and I'm flexible. Mike, to your comment, I had the same situation. So when I lived in Bermuda, you had a choice. You either got Bermuda holidays or you got a us holidays and every employee made a choice and they are different. And the key is, it was just like to let everybody, you know, you just had to communicate to people, Hey, this is your holiday. And, you know, this is my holiday and, you know, we'll work around it. And, um, fortunately it's the odd balls where, where you run into trouble. Like, like I would say like today, the jubilee is going on and we just had Memorial day in the United States, but the summers tend to be okay. And like in Bermuda, they had, um, cricket holiday on August 1st and second, the entire island shut down for two days so that, so the North could play the South in cricket. Um, and you know, that was one of the things it was like an odd, odd days, but it's the summer. So it's, so it's not as bad, but it's just really about communications and expectations. You just got to set them and expect people to live by them.

Async tools and documentation

Aaron, I see you asked a question in the chat. You want to ask live?

Um, on async work, like I guess being productive with others, especially when conversations get complex, maybe there is no good tool other than a meeting or a call like email versus Slack versus mirror. Like, are there any async tools that you like to actually work through something complicated and make decisions without the, without the meeting?

That's such a good, I'd love to learn from others on this one. Myself. Um, I will say what, what we do. I mean, we try to lean into GitHub whenever we can, you know, I think the process of writing down where you're stuck, um, and you know, like here, here's my reprex or what I'm doing. How do we solve it? Like, that's good when working with non-technical, I mean, like, you know, non-data scientists, um, you're right. Sometimes you just gotta resort to the meeting and that's, and that's fine. Um, it's, it can be a lot more efficient that way, but, um, the, the drawback there is you don't end up documenting. I think a lot of us have learned in the last couple of years with the remote setups that, that really just the writing things down reduces those institutional knowledge silos, um, where you can never pick up if someone leaves and go somewhere else and you go, Oh my God, we don't know how to do this anymore. Um, instead of the meetings whenever possible, writing things down is really wise, but as far as tooling, I, another one I wish I had a good answer for like, Oh, we use this consolidated approach with this thing, but now we're kind of spattered all over the place. We have some Slack, some teams, some GitHub emails. Um, yeah, I don't know. Cause I, you know, I, I was, um, interacting with someone on tour recently about loom, uh, and that looks awesome to sort of share it. So loom is a platform for sharing quick little videos, um, screen shares with like, I think it has like a little headshot of you talking in it or something like that, and it seems like a really great solution for asking and answering technical questions and in a streamlined way, but I have not used it myself. Curious if others have recommendations there.

So I've used it before for async work, especially with project kickoffs, when you can set context really easily with a short video, you can reduce meetings, get everyone on the same page and then have everyone kind of digest all the information before you meet for more technical work. So I found that's really important whenever you're kicking off projects. Um, also for updates that you maybe would have a weekly sync on. Sometimes you can just do it in four minutes over a loom video. And then if there are any follow-up questions, someone could send you a loom video right back.

Remote work: pros, cons, and the pandemic shift

Travis, thanks for speaking with us today. Um, I wanted to ask you about, um, your ability to work remotely. So is this something that you've always done or is it something that became possible because of the pandemic? And I'm also wondering, you know, I hear, I always hear a lot about how there's a lot of new roles that are opening that are remote, but I'm still curious, is this like, are people hiring, um, with remote roles in mind for data science, or is this still something that needs to be like negotiated? Um, like pretty intensely in order to have, cause I know there's a lot of talk now about return to work and, um, a lot of CEOs are saying people need to come back, so I'm just curious, um, how this became possible for you and if you have any insight into what it maybe looks like now in the market.

Yeah, sure. Um, so for like a couple of things there, that awesome question. I love it. The return to work phrase is like usually a hot button, um, phrase with huge remote proponents. Um, so it implies sort of that we weren't working before, like when we aren't in the office. Um, I know that's not what you meant, but, um, but it's just sort of a funny thing that, that it, that it struck me. Um, but yeah, I think if it. In, in the negotiation stage, which you mentioned, um, so if you're looking at a role and, and it turns into a company or an org saying, well, we're going to let you work remote, um, as if it's some sort of privilege, it could be a red flag if, if remote really is your intention and your desired mode of working, and if you work well that way, it's just the way you work. I mean, I don't, I don't think it's, it's a, it's a privilege or not. It's just, it's the way that you're productive. And if, and if orgs don't see it that way, then, then maybe that's a, that's kind of an early indicator.

Um, so it's an important conversation to have early on, you know, like when, when we're interviewing new people for, for our team, we make sure to say like our org is a hundred percent remote, um, and indeed for us, and even for me, the a hundred percent remote, like going away remote only became possible because of the pandemic. Um, so kind of one of the few bright points of the pandemic, it allowed us to say, we really can do this from anywhere at any time, um, but now that we've done that and it's successful and there are of course, cost benefits to not having brick and mortar, uh, all those things, it makes sense for us. So we say like, you know, our intention is to stay a hundred percent remote and that's what this role is. And then there's no, there's no kind of like, we're gifting you this, this benefit. Um, it is just the way it is. Um, cause people are, I mean, a lot of studies I think lately are showing that people even became more productive when, when they work remotely, which makes a lot of sense to me. I know I did too. I don't waste two hours a day sitting in a car and, you know, putting on proper pants or whatever the case may be.

Um, so it's an important conversation to have early on, you know, like when, when we're interviewing new people for, for our team, we make sure to say like our org is a hundred percent remote, um, and indeed for us, and even for me, the a hundred percent remote, like going away remote only became possible because of the pandemic.

I definitely, um, I'm sorry. I definitely meant return to, uh, return to office, not return to work. Yeah, yeah, no, no, no, no. I didn't, I didn't mean to sound accused. No, I know exactly what you meant. I didn't, I didn't sense any like ill will there. Um, but I definitely think you're right. I think that this pandemic has definitely shown that there are some things for data scientists that are just, I know I've personally found easier to do at home than in the office. Um, and I think this conversation we're having today is really bringing up important points about maybe some of the challenges with, with remote work. Um, I know we're talking about loom and how that can be really helpful. I find another thing that still maybe I find more helpful in the office is brainstorming. And the main reason is because of whiteboarding it's, um, it's really hard to whiteboard over zoom and like brainstorm those ideas.

Clinical trial phases explained

Um, but in simple terms, what is the difference between early phase and late stage clinical trials? What's the difference in the work there?

Oh, yeah. Um, sorry if I, I started with, if I started with jargon, um, this whole thing, that was not what I meant to do. Um, so the, um, clinical trials as a whole, uh, have phase one, two, three, and sometimes four. Um, so phase one and two are kind of early investigational studies where you're looking at a new drug and you want to know, is it safe? And do we have any hints that it's going to be effective, um, for a particular disease or patient population and then later stage trials, um, or, uh, phase stage, um, the, the three, four. Um, so three is the, the classic randomized controlled trial where, where you're saying we believe it's safe. We believe it's effective. Let's prove it. Um, and then a four is usually a very large population wide. Um, you kind of surveil the, the, you know, like lots of people are taking statins these days. You could run a phase four or something like a statin and make sure that it's doing what it's supposed to do and that it's still safe and, and, um, larger populations, sometimes even harder to reach, um, aspects of the population that might not have enrolled in the classic phase three RCT.

Um, and then, uh, I was, I was referring to late stage, um, prostate cancer. Um, prostate cancer is a disease in which, um, a lot of men are diagnosed very early on, so an early stage of disease, um, and the, the kind of actions you would take clinically for early stage prostate cancer are very different than those you would take for late stage prostate cancer, which is most often test the size to the bone or other places. So we mostly operate in the, in the late stage, um, phase with, um, metastatic disease, um, kind of low life expectancy and, and a real opportune time to introduce new drugs that are helpful.

Convincing leadership on remote work

Um, but it was, how do you convince leadership in pharma companies that are still hesitant or resistant toward remote work, that it's the future for data science?

Um, oh, that's like the million dollar question, I guess. I, um, I suppose I've been fortunate not to be in that role. I've had to make that, I've not been in a position where I've had to make that argument myself. Um, so I'm gonna, I'm gonna cast an, I don't know, into the, into the, um, into the chat, uh, if anyone has suggestions, I'm, I'm really interested to hear them. I read a lot of blogs and a lot of kind of like Twitter takes on this. Um, and you know, like some of the hot takes are just sort of like, well, I'll leave, you know, we'll, we'll leave and go get remote jobs that make more money. And I mean, like, sure. Um, but if you like the work you're doing and the place you're, you're working, that's probably not the best solution. Um, if you'd like to, you know, have a civil, productive conversation.

So in, um, in my experience, kind of in a previous role, um, we did have a group that we're working together, um, in one location and in that sense, we, there were certain individuals, this is pre pandemic, but certain individuals can, in order to get work-life balance and right, and, you know, kind of get that in place where they were having to commute a fair distance and take kids to school. And so their, their work day was kind of difficult because of all of that. They, they were trying to make the point that perhaps it would be more productive for them to be at home certain days in the week so that they could then, you know, get the kids off to school, do a full day's work, go back, pick them up. And deal with everything. The line manager at the time was kind of saying, well, because they wanted to keep a cohesive group present in that location, if individuals were kind of off working in remote locations, that was harder. I can kind of understand that rationale just in terms of a group dynamic and the kind of over the coffee pot, you know, round the kettle kind of discussions that you occasionally have with colleagues, or, you know, you just turn around in your chair and ask questions, that's much harder when everyone's remote. So I kind of understood that point of view.

I think, well, just before the pandemic, I made the case to work remote and, you know, my manager at the time was like, sure, you know, you're, you're on the phone 90% of the time to folks in a remote location. So why should it matter where you work? And then the pandemic happened. And as far as I'm, as far as I'm aware, I don't think that productivity has markedly dropped within my organisation in the last two years. So I think it's a much stronger lever now to say, hey, look, we can do this remotely. Sure, there are some jobs that are more tricky, or certain circumstances that are, you know, maybe you need to have closer interactions with colleagues. Like if someone is brand new into the industry from graduating, you know, you maybe want to have closer interactions with them to make sure that they're getting it and they're not kind of, you know, veering off the path too much, let's, let's call it that. But I honestly think that once you kind of get to a level of knowing what you're doing in your job, I can't see why anyone could have a strong argument that you have to be in the office.

Yeah, that's a great, I love that story in that you're pushing for pre-pandemic. And I think for a lot of places, the pandemic just was the proof that said, look, we really can do this. And like, you know, it's, it's, it's almost like the onus is now on the, the in-office people like, well, show me how productivity increases when we're in the office. Cause we've, we've, we've proved it like in, in the most dire of circumstances, like in a time when we shouldn't have been productive. I mean, that's the thing that the COVID enforced remote style of remote work is not normal remote work. That's not the way that it's supposed to be. It's supposed to be more of a choice and a, you know, an intentional thought process than rather than a, than an emergency knee-jerk reaction. So that we did so many companies that so well remotely in, in the worst of situations means that they'll do even better in normal times.

I mean, I, I have to say that I think communication is much more difficult. You know, you have to be intentional about communication. And also I think that it's easy for individuals to feel isolated, you know, and that's, that's got a knock-on effect on mental health and everything else. So, you know, I, there are pros and cons for sure. But, you know, I think having those honest discussions about them is, is good first step.

Yeah, totally agree. I, something I meant to mention about remote work in general, I learned a lot from the GitLab suite of resources. They have so much out there. I mean, they have a really good Coursera course. I think it's called remote team management or something like this, which is very good. And then they have a remote team handbook, which speaks to a lot of those topics. You know, how do we, they always say the biggest, you know, challenge is communication and, you know, document everything, but that's really hard. And how do you carve time for it and make sure it's done? And how do we balance just kind of connectedness in the long run? Certainly things to think about and be intentional about.

Staying connected with remote colleagues

I know on this topic as well, Ethan, you just put a comment and question around like staying connected to your coworkers. Do you want to jump in and add some context?

Yeah. I was just wondering kind of what Mike mentioned, I think about people feeling more isolated. I know I've felt much more isolated. I've, I've worked now at two companies, neither of which was big on remote before the pandemic and maybe haven't put a lot of thought into like doing remote well. They've one of them, the one I'm with now hired me on fully remote, which was wonderful. Um, but, um, I definitely miss more of the interactions with coworkers in person. Um, but they're all in Ohio. I'm in Mississippi, so we can't like just meet up. Um, and I'm just wondering what, uh, other people have seen that has worked for helping you stay connected.

Do you do any of you all work? I was wondering if any of you all work at the companies that have done the, the, um, the remote, like if you're a fully remote company, something that's become, uh, hip maybe these days is like the retreat sort of idea where you just go to like the place in the country or somewhere else and just go, um, we haven't done that at our org, we've, we've talked about it a couple of times, but I'm curious if anyone's done that and if it was awesome.

Actually, I want to say like a remote job is good. I completely agree with Mike also, because I have done remote job also and work from office. So both experience I'm having, when I was a fresher pass out from the college, then the remote job was quite. Very complicated. I will be honest because I don't have any experience how the IT industry work or what is the domain they are working. So in sitting in isolated room and in the front of computer for nine hours, it's like become frustrating. Your language is telling you like, do this by that way. Like that is quite fast for a beginner in the industry, but, uh, in the office environment, the most important, I think I love is that interaction with the people you are hanging out, you're working. If you have a problem, then you are referring, Hey, how do you do that? He will tell you it's so simple way, but in our remote, because at present I'm working remote as well as office. My team is like in three or four nation, Australia, UK, Abu Dhabi and India. Like my 11 team members are divided into the four nation. So we are handling the project of stock market domain. So the interaction is quite good, but sometime it's quite complicated also because, uh, first of all, the language barrier is not there, but pronunciation barrier comes around how you're understanding your mate, like, because you don't see their facial expression. You are just hearing your voice. You have never met them. And like, so I would suggest that working from office is much and much an interactive, funny, and like nine to six hours, my work time was like. So fast that I can't believe that. And nine to six hours in the front of laptop in my home is like a 24 hour job.

Um, I actually, a couple of weeks ago, I got back from a retreat, um, with my company, we're fully remote, we got about 40 to 50, um, employees and, um, it was really fun, uh, there was a little bit of work, you know, sort of during the day. But then, um, you know, like in the afternoon we like did a scavenger hunt around the city and, um, or we did, uh, I mean, all sorts of just silly little activities and they were like, you can go to this mall or you can go, you know, to this museum or, or whatever. There were just all sorts of fun activities. And then in the evenings, of course, there were dinner and drinks and everybody went out after the dinner and the drinks and had more drinks. And so it was a, it was a very fun, a very fun time. And it was really fun getting to see everybody. Um, but unfortunately it was a COVID super spreader event, but I was boosted, so I'm, I'm good, but yeah, it was fun. I would, I would love to chat more about it offline if anybody's interested.

Yeah, that's really cool. Sounds great. Minus the COVID part. Yeah, exactly. I saw, I saw someone actually mentioned something that I also, um, do from time to time where there's this, um, you can co-work with your, you know, like in a, in a zoom or any kind of, um, teams meeting even silently, um, especially with the onboarding process, I find that's really helpful so that, that the person or people know that I'm there and available, um, like I have things to do, so, I mean, I'll just sometimes even turn off our camera and just like go on mute and then, you know, if you want to like pop on and say, Oh, I got this thing. I'm stuck. I got a question. It works really well. You can block an hour or two with, with, um, certain colleagues who do that from time to time. It's kind of like working side by side.

Python vs R in clinical trials

There's a few other questions that were on Slido anonymously that are in a different topic and one was, um, does Python have as big of a place as R in clinical trials?

I've, I haven't seen it yet. Um, I, I could be totally blindsided by something that I'm just not aware of. Um, but I, I think the reason is that R was traditionally built and is more maintained in the classical statistical space. Um, and there's not too much NLAI, which of course is a, is a Python strong point, um, going on in the traditional clinical trial world. So, um, just the strength in terms of you're going to run a regression or a survival analysis and you're going to get exactly what you're going to get. Um, I think that's sort of accepted as, as a more, as a truth, um, compared to, to Python and I, and this is not a slight towards Python. Um, like it just, I just think it's the different way that these languages have evolved.

Commenting code and code reviews

I think Javier, you asked the question around, um, like commenting in your code. Would you want to jump in and pass that one?

Yeah, I, I don't know if I'm unique to this because now at, um, bloom reach the company I'm working at now, you know, our team is much smaller. So I, over the last few years, I don't know if it's COVID related or not, but I have been putting, you know, a lot less comments in my code. Um, I think part of that too, is like our teams have just sort of matured. So there's less of a need to comment stuff that, you know, is like clearly obvious what's going on, you know? Um, do you have any advice for like when to draw the line of, I should comment or put some pros in my code here or not?

Um, the, so I, I'm guilty of this. Um, I've definitely written like some code, especially when I, so I, I started the data science team, um, that, that I'm leading now at this org. So when it, when, when it was initiated, it was just me, like writing a ton of code and, um, I didn't always comment that well, which was not wise for future me and certainly not wise for, um, other people that would be joining the team. Um, what's, what's been helpful is, um, I work for example, with, um, Shannon who's awesome at communicating with me. Uh, this was hard to read. Um, so I, like, I honestly lean on others to tell me what that line should be. Um, it's something that I think it was Alison Hill might've mentioned today or taught me about this, which is like a, almost like a feelings journal for, for how you're experiencing a set of code, like when you open up the R document, if you just kept like an informal side document, which is when I saw this line, I felt, um, happy, um, because it was well commented or I felt like what is going on, um, like those sorts of things, feedback from colleagues when they're, you know, you do paired programming or code review is really, really helpful to me to know, like something that just might be clear to me because I've written the statement a million times is like not super clear without a comment.

Going fully remote: tips and dedicated workspace

Yeah, I think, um, for a lot of us who have had the more traditional in office environment, um, and with remote work becoming more, um, popular, you kind of get a grass is greener on the other side kind of feeling, and you're thinking maybe I do want to go fully remote. So I think I'd just love to ask the group, what are maybe some pros that we should consider, but also some maybe downsides or cons that someone should also consider before thinking about that switch from being maybe partially remote or fully in office to being fully remote.

A quick take for me, um, would be a couple of the, I mean, the benefits to me are such that I'm here. Um, so I'm clearly a biased observer here, but, um, of course, some of the challenges we've talked about are the communication and maybe potential isolation. I mean, that, that really is something that. You, the, the team has to be so, so thoughtful about to make it work well and make sure that everyone's still engaged and feels like they're part of something that's meaningful and that everyone's growing. Um, that's, that's something I really try to keep tabs on in the team that everyone, you know, like, did you learn something this week or, you know, how do you feel about your future and like next week and, and not, because it's, it's very easy, I think in this potentially isolating environment to, to just have, like be in the grind of like producing the thing, like, you know, Oh, you did your tables this week. Great. Like let's move on and we'll do more tables next week or whatever the case may be. Um, just checking in about people, um, and, and making sure that, that they're balanced and happy is, is something if it's possible to probe early on, like what, if you can get a vibe, um, I'm sorry that there's not a more rigorous thing to look for than that, but, um, that would be my two cents.

Yeah. So I, when I first switched to remote work, I was working, um, for the federal government and, um, I hated it at first, um, because, you know, I had this nice little commute on the Metro and I would, you know, do out and go out to drinks and everything afterwards, you know? And so I would, so I hated, hated, hated working from home. Um, but now I love it and I won't ever do a job that's not working from again. I think my biggest piece of advice, if you are looking to switch to fully remote work is to definitely get your, like a workspace, like a total, and that's what you do in that space as you work, because when I first started working, my workspace was also in my bedroom and that was just like such a bad vibe generally. Um, so yeah, definitely. If you want to make the switch, make sure you have a dedicated, and it doesn't have to be like a whole room. Like I happen to have an office right now, but I think if, as long as you have like a desk or something, that's like a dedicated workspace, um, I think that's really important.

Totally endorse that one too. Yeah. Sometimes when the weather's bad, like I feel that when I work in my room, like in the RV, which is, you can imagine it's actually pretty small, um, that, that feels very crushing, um, like soul crushing. So I have a tent, um, that I, that I work in when the weather's not great. Um, and I have like a little space heater that I'll put in and things like that. And even just that simple walk for my RV to the tent is like, okay, I'm going to work now. Um, and so, yeah, I completely agree. The dedicated workspace is huge.

Pair programming and knowledge sharing

Um, Eric, I see, Eric, you had a question earlier on commenting code. Do you want to jump in?

Yeah, definitely. Um, so I, like others have been saying in the chat, thinking of future, you is quite important when you're commenting on code and the like, but I'm one effort I did earlier this year is it's not just about the code for a specific project, it's how you're accessing certain environments or maybe using reusable, you know, code across projects or just other kind of intangibles. I'm just curious how your team is approaching, how you document and share your learning about, you know, how you're accessing certain tools in your infrastructure or how you're trying out new, you know, projects that you're trying out new, you know, ways of doing analysis and how you're sharing that learning across so that nobody's like trying to do it all themselves every time, trying to avoid that syndrome.

Yeah, that's, that's such an awesome question. And something that I, that I worried about, not on this role, but in my, in my last work too, how do we create like a wiki that is just, suppose you landed in like a data scientist role here tomorrow, like, how would you know? Like what, where are the keys? What are the, what are you, what are you doing? And I, I, something I pursued in the past was like GitHub pages via Netlify, so making just like a simple, like R Markdown site and, and putting that together, the, the limitation that I found there was then the, the people who can contribute knowledge to those pages are only those who know R Markdown. And, and that's big because I want people who don't code to feel like they can come and comment on our processes and, and, and share their knowledge as well. So these days, as, as much as I'm reluctant to say it among the data science crowd, we're just using like SharePoint, like, which is fine. It's easy. It's lightweight. Like, I, you know, the, the, it's kind of already licensed through our org. So it was free. But other suggestions, I mean, any other thing that people are using there would be, I'd love to learn about that as well. What, what's in, what's a low barrier to entry Wikipedia knowledge base kind of toolkit? I, I haven't stumbled on my favorite.

I can tell you that what I did earlier this year is I did do an R Markdown based site with blog down, but I found a theme where most of the content is just Markdown itself. It's not like I'm doing R specific Markdown syntax and I was able to tie with GitHub action so that somebody could go on a page. They would see a button that says edit this page. If they see something wrong or I've seen improvement, it would take them straight to GitHub in our org for that page. I could just edit it on the spot and GitHub itself, and then put the poll request, I merge it in, then there's a GitHub action that redeploys it. Um, that's very cool. Yeah, it took a lot of work to get there because infrastructure at big orgs is always a complicated mess sometimes. Once I got a little help from our IT group on setting up the deployment piece, I'm much happier with this solution than what I was doing in the past, which was in essence, scattered files on a shared Linux, like file system space that nobody could index or find or anything like that. So it's been helping future me at least.

I see Raha, you asked the question around code reviews too. Do you want to jump in?

Yeah, sure. So right now in my group, the only person coding is really me, but I'm going to start with a, I'm hiring a postdoc and then we're going to bring on a bunch of undergrads and I would like to, I would like some sort of mutually beneficial relationship where I can give input on how their coach could be structured better. But I also need to have them look at mine and all of the things I didn't comment that I should have. And I don't know what the best way to implement this would be also taking into account that for some people, their code might be like, it's, they might think of it as their property. Like they don't want to give it away to another student who would just recreate what they've done. Um, so I want to protect them, but also make them better.

Yeah, that's a good question. So this is, um, Garrick Todd. I think I see Garrick, you're hey, you taught me about this. Um, like pair programming strategies. Um, something that I found is, it's hard for me is when someone says, can you review my code and then it's like points me to a GitHub repo, and then I don't know where to start and it feels like a lot instead, um, pair programming strategies whereby maybe one that Garrick was teaching about is, um, a driver passenger sort of setup where someone codes and someone else is kind of like the passenger and saying, what are you thinking now? Like, what are you doing next? Um, and then you kind of iterate together through, through the code base in real time as you develop it potentially. So it's, it's lesser review like post-mortem, um, and more of a, like, how do we, how do we think through and build? And that's, and that's really good for, um, onboarding people too. Um, but then that, that could also work for simple code review if it's already done, um, just kind of like looking block by block, you know, like have maybe the, the person who wrote the code could be the driver then, and then they go, okay, here's the first thing I was going to do. I opened this file and then I was going to load this data and here's my thought process. And I think it kind of, I think it really speeds up the, the workflow instead of just like feeling lost when you land in a, in a big code base, um, which is usually my experience.

So I have the pleasure of being probably one of the very few, like handful, maybe two handfuls of people on my like 80, 90 person team who uses R. And so I noticed one of the challenges is if I wanted to do pair programming with someone, um, a lot of people actually don't know the language. So I don't know if it's just, you have to find someone else who knows it and try to do it only with them, or if anybody has tips around what to do in that scenario.

That was actually how we stumbled on pair programming originally. Um, Garrick and I was that we were trying to onboard people who did not know R. Um, so it certainly can work there and you're right. I mean, it kind of limits, like you sort of, in many cases need to be the driver if the, if the person doesn't know any R whatsoever, but it can be valuable for, for kind of, um, more entry-level tasks if, if that, the, you know, the learner is then the driver and then, you know, make clear that we have all the time in the world here, like, it's okay to not know what to do next. And just, you know, like lean on me when you need it. Um, that, that works too. Oh, Garrick. Hey, do you have, do you have other thoughts? You've done this more than me. So I feel like I'm just trying to channel you.

Yeah. Awesome. I mean, just talk about pair programming. I, um, I love