General Presentation Tips Video Lecture Transcript
This transcript was automatically generated, so there may be discrepancies between the video and the text.

Hi, everybody. Welcome back in this video. We're gonna talk about some presentation tips and tricks with some general presentation tips. So this is just gonna be a notebook where we go through and I'm gonna give some general tips for giving a data science presentation. Uh These are gonna be primarily geared towards your pre-recorded video presentations if you're currently working through a boot camp.
Um The while that is being well, that's uh with that said, uh this, these are just general tips that should probably apply to most presentations you're given. Uh So over the past few years of a code of doing these coding boot camps for the Air Institute, we've watched a lot of these pre-recorded video presentations and so these tips sort of encode useful information that we've gathered from watching these videos and seeing what works and what does not work in a, in a data science
presentation in this setting. Um So the first tip I want to talk about involves time limits. So think back to your entire life as a, as a student and as a consumer of presentations in various settings. Uh certainly there has been times where that presentation or that class had its preset time limit. But the presenter, teacher, lecturer, professor went over that time now, sometimes, maybe this was a really interesting and great presentation or class and you were happy to have
more of it, sort of like a little bonus. Uh Other times though, uh maybe this was a really big inconvenience for you or for your fellow classmates or presentation attendees. Uh Maybe because it cuts into what you already had planned for the rest of your day or made you late for the next meeting or presentation. So, respecting a time limit of a presentation or a lecture is very uh important because it means that you're respecting the time of your audience who without them, there'd be
no point in making a presentation. So the key tip here is if your presentation has a set time limit, you need to make sure that it's either at that time limit or slightly under. Now, it's not to say that you should never go over in a presentation, use your use your best judgment, but a really nice general rule of thumb is this, you know, to make sure your acts were slightly under your time limit.
So here's an example, if you thinking of your five minute pre-recorded video presentation, here's some acceptable video links. Uh So four minutes, 45 seconds, four minutes, 53 seconds, four minutes, 57 5 minutes. Uh sometimes you may go a little bit over just because you just can't trim the, a few seconds, but something like five oh three or 5 10, these are both acceptable.
You're still, you're going over that five minutes, but it's not by too much. And I would say that like, in, for my personal preference, like 10 seconds over is probably about where I'm starting to reach the limit of how much I'm ok with you going over. Um, so I would stick to definitely under 5 10. Just that, you know, using this extra time that goes over is like, I can't cut anything else out of this presentation.
Um If there's something in your presentation, you can cut that would take it under this time limit, you'd need to do that. But if, if everything that's in there is like when the, the information your audience needs, keep it in, it's ok to go a little like 3 to 10 seconds over not more than 10 seconds. So video links that you should avoid are anything beyond that.
So like 5 35 47 5 55 6 30 in the past, we've even had some people that had videos that were over 10 minutes for a five minute time limit. So that's twice the size of the, of the time limit, right? So just remember remember that this is all about being respectful of your audience's time. The audience has come to watch your talk. They want a good talk, but once you start to push the, the limits of what they're able to give you, uh, they're not gonna be, you can really break the, really break the
great presentation that you've given up to that point. So make sure to be mindful of your time limit, try and stay slightly under or at that time limit with the note, you know, it's, it, it's ok if you go a teeny tiny bit over, uh, but try and stay at the time. This is probably my biggest tip of the entire notebook is respect your time limit ... with that in mind.
Uh What should you do within that time? You need to tell a good story. So good presentations often tell a good story. This is a nice piece of advice that my phd thesis advisor gave me before my first presentation as he was helping me develop it first academic presentation, not first presentation ever. Um but it's about telling a good story. And so what I mean by this isn't like you need to be the world's greatest orator that's telling a tale as Timeless as the Iliad or the Odyssey.
It really just refers to your ability to clearly and concisely confer what your project or presentation is about to your audience. So in my experience, these are the things that I found uh are key to making a good story out of a data science or research project. You need to have a clear problem statement So think about what do you want your audience to take away from this presentation?
Make sure that is exceedingly clear, like hit them over the head with it. Uh Don't make it a subtle point. Don't try and play like a an M night Shaalan game here, make it a very clear statement like this is what this is about and I'm gonna tell you about it. So think of this in some sense as like an elevator pitch for your product. And if you've never heard of the term elevator pitch, it's sort of taking a short 1 to 2 sentence description that would you would be able to tell somebody while
you're riding on an elevator with them from maybe the fifth floor down to the first floor, right? So think of this as you have a 1 to 2, maybe three sentences to explain like what is the essence of this presentation and this problem that you're working on and you should confer it. It is typical to go along with that problem statement to maybe provide some motivation.
So something like this is the problem that we are going to work on and potentially hopefully solve in this presentation here is why this is an important problem because of X Y and Z. So your motivation should include importantly, an identification of who cares about this problem beyond the audience who may not have a stake in the problem directly, who are the people that are most affected by the problem that you're given, uh, you're, you're talking about and who would most care
about the proposed solution. So, uh, maybe it's just like here is what I'm working on. Millions of these types of people, doesn't have to be millions. That's an exaggeration. But, you know, this large group of people are affected by this every year. And if we can solve this problem, then this will be the direct benefit to them. Something like that doesn't have to be exactly like that.
You can still work on problems that don't impact a large group of people. Um But that's just an example. If your time permits, you can include some slight background on what you're talking about. Uh maybe stuff that's been tried in the past and how you might improve upon that. Um It just depends again on your time limit. So the next thing you should talk about is what you did or what you tried to do.
So think of this as like the methods of your, of your talk. So you, you've, you've laid out what is important in my presentation, what this is gonna be about what I'm trying to solve. Here's how I'm gonna try and solve it. How much detail you go into depends upon two things, uh your time limit and your audience. So for instance, if you're a think about this in terms of presenting your research for your academic research to somebody, um if you were to go and present your research to uh and
I, and I might be actually giving this later. No, I think I talk about it here. Uh OK. ... No, this is, this is not where I talk about. So we'll come back to that analogy in a second. Uh So some common topics to where you're like gonna talk about what you're trying to do. Maybe you want to explain a complicated data set, uh uh data set and what you needed to collect to get it, what modeling approaches you might be taking.
And you remember, you might not wanna go into super deep detail depending on your audience. Uh And you know, how did you decide to gauge model performance once you were done with those models? I think an important facet of trying to explain what you did is thinking about the problem statement and what stakeholder like how did how they affect the stakeholder?
So for instance, like when thinking about your performance measurements, why did you choose that? Usually you want it to translate it into some sort of impact? So for instance, uh I chose to focus on the true positive rate for this classifier because we wanted to make sure it captured all of the positive cases of cancer that exist uh in our, in our sample, something like that.
Uh So then after you talk about what you tried to do, you need to tell them. So what actually happened when you tried to do that. So I think um the specifics of this really depend upon your audience. So how deep you dive into like what you found with your modeling? So I think one way to think about it is um ... well, actually, I keep saying I'm gonna go and my audience comes later.
OK. So, you know, it just depends upon your audience. I wanna say, you know, this section doesn't only have to include the successes of your project. So if you had a project that is like full of Chaka block, full of 100% success, that's great. But oftentimes along the path of your project, you are gonna encounter difficulties. This is an OK time.
Uh If you have the time remaining here in presentation to talk about out what some of those difficulties were because you can use it to pivot in the next section to say, hey, we encountered these difficulties. If I were to continue working on this project, here's what I would do to address them. So they will no longer be difficulties as I build this up into a bigger deal.
So uh if you're thinking about this, like in academic paper, you know, this first part was our introduction. Second part was our methods, this is sort of our results. Uh The next thing after your results would be you want to make some comparisons to what's been done like. So if it's applicable, so let's say you had a model that people have worked on this problem before, but now you have a new model that improves upon, you know, the standard performance by X percent or something like
that. Or maybe like you're presenting a new approach, which maybe it doesn't improve accuracy or something like that, but you're, you're improving the time at which results come out, something like this. Um If that's applicable, you might want to include it. You may also want to talk about what are some future directions if you were to continue working on this project.
So what would you do next? Uh A lot of times these projects are done in a time crunch. So it might be useful to talk about. Uh This is where we are so far. The next thing we need to do is this or the next thing we need to do is go back and set up a cleaner uh data collection and cleaning pipeline. So that way the results are easier to get to something like that.
And then at the end, you need to have a good summary and conclusion. So this is sort of the last thing that people are gonna see from your presentation that's substantive. Uh So this is, you know, remind them what your problem statement was, remind them what you did, remind them what you accomplished and then have a nice summary conclusion sentence.
Um So this is you want this to be sort of like the takeaway message of your presentation should be encoded into this last section. And then at the very end, typically, after you're done presenting the important material, like of what happened, you also should give recognition to the people and groups that deserve it. So for instance, uh if somebody was very helpful to your group, you want to put their name here, like if you had a mentor that wouldn't actually do it, the workbook
gave you some nice guidance. Uh If you're part of a research team and you're representing that team on your own, uh If you have some sort of funding source that provided funding for this project, all of this should go here in a single slide. Doesn't have to take too long, but thank the people and groups that deserve it. OK. So here's the uh second step or third piece of advice that I kept wanting to get to uh in the earlier section, know your audience.
So it's really important you should frame your presentation. So it's at an appropriate level and engaging for the audience that you have. So one way I like to think about it in terms of academic research presentations, which I'm sure most of the people watching this video have experience with is think about this sort of scenario. How would you present your research results to these two groups, a conference room of researchers actively working in your field.
So experts in this field versus an undergraduate introduction to research seminar. So typically, maybe this is a seminar undergraduates have to take because they're interested in your field, but don't really know a lot about it. Or maybe they just have to take it because it's a requirement from the university. When you're talking to the first group of experts, you may be more likely to go into detail about specific techniques and innovations that you're doing in your work.
Whereas to an undergraduate group, it might be better to give like a broader picture of the field as well as like what you did. So this is what you do for academics. This is entirely familiar to all of us or most of us, if not all of us, the same settings, uh same steps hold for industry settings. So if you're giving a presentation to an in an industry setting to a group of data scientists that are working on the same problem as you, you might be more likely to dive into the details and
present it in a more technical manner. By contrast, maybe you're presenting to upper level management that isn't as down into the details and maybe is not as technical or maybe you're just pre presenting to other team members in the uh in the company that you're working in, that aren't technical members, but have to go out and translate what you're doing in the technical team to attract clients and customers uh to the product or, or uh uh service that you are providing on the
technical team. So just be aware of who your audience is and let that determine how much detail you go into on the nitty gritty stuff. ... So here are just some, these like first sets of tips are really just like broad presentation tips. Now we're gonna get into sort of details of like things that you should do in your slides themselves. So when you're doing slides, be clear and concise.
And what I mean by this is like a presentation isn't the same as a paper or a book where you sit down and you write full sentences and paragraphs with really good grammar. Um You don't need to use full sentences and I think ideally you should avoid any paragraphs of text. So one of the things that I think turns people off uh in presentations that I've seen a lot in these videos is people will put up a slide that's a high, a huge wall of text that maybe is multiple paragraphs long and then in a
setting where they have a short time and let's spend like 20 seconds on it and then move on to the next slide. In most cases, people are not able to both read a wall of text and listen and process what you're speaking. A lot of people can't multitask in that way. So it's up to you to find a really good balance of clearly conveying your ideas in your slides uh with the content.
Um So maybe that involves writing slightly more detailed sentences, uh but also conciseness. So, is there a way you can take that big paragraph and, and boil it down to five bullet points, something like that? Uh A big pet peeve of mine are, are presentations where the plots aren't very good and by very good, I mean, like it's hard for me to see them and understand what's going on.
Not necessarily that I think they're bad plots like in terms of their construction, but just like the physical ability of me being able to look at the screen and, and read the plot. So I think it's important to put in very visible non blurry plots or tables. So these mean your plots and tables should be clearly labeled, they shouldn't be blurry or like stretched or um disfigured.
Uh I want you, I would prefer to have really easy to read tick labels. If the tick labels are conveying information, sometimes they're just there. Uh And then, you know, try not to distort it. So sometimes what can happen is if you're copying and pasting an image over, you make it bigger by stretching it out. Uh And that distorts it makes it kind of grainy and hard to read.
So in the, in the second notebook, in this uh presentation tips and tricks little section of our content, uh There's a nice plotting tips where we can go over some steps to, to avoid these sorts of um uh these sorts of errors. Uh I also think you wanna be able to make sure you include all relevant information. And so I don't mean this in terms of relevant information for your project, but relevant information about like who worked on the project.
If someone wanted to find your slides and they're available, where can they find them, where can people find any supplementary materials regarding your project? So for example, let's say your project has a github repository that stores all the code and shows people how that code works. You wanna be sure to include that link on the first and last slides.
I wanna say importantly, these links should have the full text for the entire thing. So not just like a little icon, uh you should put in like w w w dot github dot com slash repository name and you should put that on both the first and the last slide. So the really re the, the good reason to do this is not everyone's gonna be able to intuit that the icon you've included, which is the github icon is clickable.
And even if they are able to intuit that maybe someone's watching a video and doesn't have your physical slides or doesn't know where to find them. So in that setting, they're not able to click on it because they're just watching a video if you were to instead put your entire full text for your link. Someone could just type it in to their browser and then get there on their own.
So make it as easy as possible for someone to be able to find the links that are important to your project. If you have links that you want people to be able to find, don't make it hidden behind an icon with no web address, do your best to make it very clear that, hey, this is some supplementary material. Here's where you can find it. And then finally, you should include all important points.
So this is slightly different than number three. So I want you to imagine a setting where you're at a talk and then someone says, well, we could go into this, but I have some supplementary material at my website. You can find it on your own time. How often are you the one that go, like maybe some of you are really good about this in practice. I'm not very good about this.
I almost never go back home or back to my hotel room or wherever. Uh uh uh you know, I'm assuming I'm in a, in an in person conference here. I almost never go back after the talk and then find that supplementary material and read through it. So the only thing that you should assume that people are gonna see is the stuff that you put in your presentation.
So if there is something that is crucial for the story of your project. You should make sure that it gets in there. So don't if it's something that's incredibly important to the project story, make sure that it's in there and that you cover it, don't relegate it to some sort of supplementary material. Most people are not gonna look at the supplementary material.
So, you know, keeping in mind that you can't put everything in a short presentation. If you and your group think that it's really important, make sure it makes it into the slides. And then my last piece of advice for everybody is you should practice. So I think there's a lot of times where we don't practice very much for presentations because we know the material.
So we think that the presentation will go just fine. But I think it's really important to not let the audience see the first time you make the presentation, meaning you should practice practicing, even if you know this material back and forth, practicing allows you to see like, what is it like to talk about this stuff for the first time? Where am I missing gaps and holes in my stories when I'm saying it with my words and it just gives you a good sense of practicing what you're actually
gonna say. And instead of just keeping it all up in your head, I think it, it's much harder when you actually have to say it than it is to think it in your head. And I think if possible, it's really helpful to practice with somebody that's not really that familiar with the project. So a lot of times when I'm practicing a presentation that's important to me, uh I'll bring my fiance in and to have her listen to me.
She knows nothing about data science, nothing about mathematics, which are the two things that I work in. Uh So it's really nice. She's able to maybe not point out problems with technical details, but it's painfully obvious if there's something that she can't follow. Uh And you know, she pays attention, she's very smart. Uh But like, it's painfully obvious to me, if there's something that she can't follow, that probably means somebody else won't be able to follow it as well.
So it's nice to practice with somebody who doesn't know the project as well as you do because sometimes our minds from knowing the project can look over missed details because we're just automatically filling it in. So the key thing is to practice and if you can practice in front of people who don't know your project as well, so like one way to think about this is they can act sort of like an editor if you're thinking about this, like a story and then they can look for the plot holes that
your current, your story currently has. OK. Uh So those are all my general presentation tips. Hopefully, you found these helpful. Uh and I can't wait to watch any presentation that you guys give that I'm able to watch. So I hope you enjoyed this. I enjoyed having you watch it and I can't wait to see it in the next video. Bye.