projects. So, when you're adding a project, there's three different ways that we can do it. The right manually, which is the way I used to do all of my projects. And but we also have two other AI features, which is an AI import and an AI guided form. We're going to focus today, just cuz we're in a time crunch to try and do this as quickly as possible, with the AI import, which basically allows you to import any sort of GitHub repos, Tableau public links, any files like Excel files or Python files or R files you've done, and write the first draft of your project for you. GitHub repos are something we can try? Oh, we got a couple that we can try. Okay, so let's go ahead and try the AI import. All right, this is Graham's GitHub uh repositories. It's definitely a little bit messy, definitely maybe needs some love, but um let's take a look. Which one of these repos do you feel like could be your first project? Which one would be good? — Let's start with the non-parametric log-linear medical costs. Okay, we'll start here. Yeah. Okay, awesome. And what is what exactly is this repo, I guess? — It's a school project that delves into different like information data to quantify uh like how smoking and different factors affect medical costs on an annual basis. — like a homework assignment? — Yes. So like anyone who has done any sort of homework assignment, this is basically just a homework assignment. Yes. And it looks like it's in Python? Yeah. Okay, interesting. I don't know much about this, so we're just going to try it. So all you have to do is grab uh the repo link right here, go back to our AI import, go ahead and give the URL. Is there any other details that we should give it or any other instructions? I don't know. All right, let's just go ahead and hit generate project article. This will take a few seconds to read through everything inside of this GitHub repo and actually do the write-up. All right, so it just finished doing your project write-up here. It made the title non-parametric and log-linear medical cost analysis. That's an interesting name. Okay, we're just going to keep it as it is. It gave you this URL slug. summary. A case study that combines non-parametric techniques and log-linear modeling to predict and interpret highly skewed medical cost data, improving forecasting robustness and interpretability. Sounds pretty professional. — That sounds very professional. And then here's an overview, the problem, the approach, data and methodology, key findings, results and impact, conclusion. All written for you. Let's look at the uh results and impact. It says more Do you remember anything about this project? — Yeah. Take a look. Can you read these results and impact and see if it makes sense or not? — Yeah. Do you want me to Yeah, read it aloud. — Okay. More reliable budgeting, improved forecasting accuracy on aggregate expenditures helps finance teams set reserve level with greater confidence. — remember that at all? Uh yes. There was like a They had like standard questions with like the data set for the like the presentation project. And there was like findings that there was like very significant correlation between like different factors and their predicted cost difference. — it's not necessarily wrong. — No. Okay. What about the Let's do this one, I guess, right here. — The log-linear coefficients in the two-part decomposition allowed me to identify which variables most strongly influenced utilization versus conditional costs, guiding targeted interventions. That's also true because they're like filtered out different factors and variables. And I think smoking was by far the like most influential factor. — Okay. So, it gets some of the results right. I guess it also said that the log of the cost was a more stable coefficient and linear relationship. So, the log was the right way to do this modeling. Okay, so this doesn't feel 100% wrong to you. — No. — Okay, cool. And it's telling you kind of an overview of the project, what the problem is, which is predicting the medical cost of something for like budgeting reasons. And then it gives you, you know, kind of how you did the data exploration, you did the transformations, and then the modeling, and then basically evaluate everything right. Okay. Very cool. So, we can hit save project right here on this project. And now if you go back to your portfolio and you hit refresh, boom! You got a project right here, right there, all ready for you to Real easy. That's what I like to hear. Okay, awesome. Let's do a another one. What is another project or another repo that we should do? Let's
do the NBA heat map. — let's do NBA heat map right here. Just going to copy and paste up here, go back to add project, AI import, paste this right here. Any other instructions? No, I think my dot MD files are pretty good. Okay, generate project article and we'll see what it does. All right, it just finished NBA shot heat map explorer. Let's see. So, I built an NBA shot heat map explorer to turn raw NBA shot and play-by-play data into actionable, visible, intuitive insights. Is that what this project's all about? — what it's about. — Okay, let's see. So, then it goes through problem, the approach, data and method methodology. So, you're getting the data from the NBA API. Correct. Okay, then doing some filtering, some special aggregation with a hex bin stuff going on. Okay, awesome. Then you're doing some kernel density estimators. Okay, great. Key findings, distinct row low profiles are clearly displayed. Hidden inefficiencies surface quickly. Strategic matchup. So, you can do team level heat maps to show. Okay, very cool. Visual artifact, very nice. So, obviously we're just pulling straight from the GitHub, right? So, it just has whatever you have in here, which I'm guessing it doesn't have like any saved images, right? — Not in that folder particularly, no. Okay, see. Well, that's something you could have told me earlier when I said, "Do you want to add anything else? " Okay, now you know. We can actually like go in and add those images as well. So, that would help you. So, let's go ahead and save project. Let's go back to our portfolio and let's get refresh on the full portfolio and boom, you got two projects. Now, I did see that on your LinkedIn the other day you had posted about this project, right? Yes. So, let's see. Here's your LinkedIn page and here's the image I saw that you posted. I'm going to actually right click on this image and I'm going to go back to our project. to the heat map explorer here. I'm going to upload that image as a cover image right here and see how it looks. Whoa. Let's go back, refresh. It looks way better. Boom, you open it up, it actually includes that image at the top now. So, I like that. Do we have any other images on your LinkedIn of this? No. Boo. Not yet? — Not yet. Okay, sweet. This is a project on our portfolio. It shows what different libraries you use in Python and obviously Python here. At the top, it will allow people to view your code and you have your full write-up down here. Um it allows people to see other projects. So, here's your other project once again. Um here's the different libraries you used here. Um and then you can always have your users go back to
