So You Think You Can ANALYZE? (Data Content Creator Hackathon)
Watch your favorite data science content creators compete for a piece of history. They will either prove their skills in front of the world or fail in the pursuit of eternal glory. This is so you think you can analyze. Special thanks to Posit for sponsoring this competition and make this all possible! Competitors: Team Jack @averysmith Jack Blandin @KeithGalli Team MMA @AlexTheAnalyst @nerdnourishment @Miki_ML Team Null Nick Singh Mark Freeman @SeattleDataGuy Team Shashank @ShashankData Team Posit-ively Skewed Greg Coquillo @datascienceharp Richad Nieves-Becker Ian Greengross Interviewer: Elijah Butler - https://www.tiktok.com/@imelijahbutler Posit Team - Michael Chow - Joe Cheng - Carolos Scheidegger - Julia Silge Sponsors, Affiliates, and Partners: - Pathrise - http://pathrise.com/KenJee | Career mentorship for job applicants (Free till you land a job) - Taro - http://jointaro.com/r/kenj308 (20% discount) | Career mentorship if you already have a job - 365 Data Science (57% discount) - https://365datascience.pxf.io/P0jbBY | Learn data science today - Interview Query (10% discount) - https://www.interviewquery.com/?ref=kenjee | Interview prep questions MORE DATA SCIENCE CONTENT HERE: My Twitter - https://twitter.com/KenJee_DS LinkedIn - https://www.linkedin.com/in/kenjee/ Kaggle - https://www.kaggle.com/kenjee Medium Articles - https://medium.com/@kenneth.b.jee Github - https://github.com/PlayingNumbers My Sports Blog -https://www.playingnumbers.com Check These Videos Out Next! My Leaderboard Project: https://www.youtube.com/watch?v=myhoWUrSP7o&ab_channel=KenJee 66 Days of Data: https://www.youtube.com/watch?v=qV_AlRwhI3I&ab_channel=KenJee How I Would Learn Data Science in 2021: https://www.youtube.com/watch?v=41Clrh6nv1s&ab_channel=KenJee My Playlists Data Science Beginners: https://www.youtube.com/playlist?list=PL2zq7klxX5ATMsmyRazei7ZXkP1GHt-vs Project From Scratch: https://www.youtube.com/watch?v=MpF9HENQjDo&list=PL2zq7klxX5ASFejJj80ob9ZAnBHdz5O1t&ab_channel=KenJee Kaggle Projects: https://www.youtube.com/playlist?list=PL2zq7klxX5AQXzNSLtc_LEKFPh2mAvHIO
image: thumbnail.jpg
Transcript#
This transcript was generated automatically and may contain errors.
Some people are doing a lot of talking and not a lot of typing. It's getting crazy in here. That right there is a hacker. I cried a lot the few days before so I have nothing left to cry. Oh it works! We're gonna win and I mean I don't even know why they're bothering.
What of the other competition has described you as a total sham and undeserving of anything good that's ever happened to you? How do you respond to that? I respect it.
Competition overview
This event will be called So You Think You Can ANALYZE. There will be five different teams. Teams of three to four people. Shashank, the previous Iron Analyst winner, he is the only one that is participating by himself. It is the privilege allotted to winning the last competition. It's a little embarrassing if you guys don't beat him working by himself but whichever team comes up with the best answers to the prompt and the data set in the allotted time of two hours and 45 minutes, you guys will have time to strategize, choose your team names, and then we will kick things off. May the odds be in your favor.
Competitor introductions
Hello everybody. My name is Alex Freeberg, better known as Alex the Analyst. I am with my team right here. We're gonna be absolutely crushing it. I don't see a possibility or any outcome where Shashank beats me in any way. That's all I've got to say.
I'm really worried about Shashank because he won last time and he crushed me and I really only came back to this trip so I could crush him. So that's my only intention and I think we are gonna do it. My game plan for the analytics is just to understand the data first. That's the first step for any data analyst role. Teams I will look out for. That guy Rashad over there, look at him. I don't trust him.
I haven't done this in months. So I'm really actually really excited to build again at this event. So whatever happens, you know, just shake out as it is. I am here to break the rules. I'm trying to figure out how do we win without actually answering these questions. Maybe making a meta tool, doing something creative. So let's see if this pays off.
I'm with NerdNourishment here in South Carolina and very excited about... What is it called again? We are going to crush it. Watch out for Team MMA.
Key strategy components are transparent, consistent communication. Making sure that we're documenting our process. So people get a real good understanding about what the day-to-day analytical tasks look like. I'm actually being used as a floater. Trying to understand each team's strategy to make sure they produce something that's really valuable for folks who want to develop insights for users. I'm just here to have a little fun. Play around with the data set and figure out if we can come up with something interesting.
We're going to win because we got, you know, positive and shiny. And we got this guy on our team. He's legend. That right there is a hacker. I'm the outsider who's going to ask all the important questions because I have no idea what's going on. And that'll help the rest of my team figure out what is going on.
What's up? Keith Gally here. I'm a nice guy. Try hard. Love the game. We've got a dream team here and there's no reason we're not going to crush this.
Early competition check-ins
Good luck, competitors. Excited to have you. There's generally two ways you could view this data. One is as a consumer of bike share or biking around. How would it be valuable to an end user? Or how could this data be useful to a business? You don't have to necessarily go down either of those routes, but framing it in one of those ways could be an effective way to communicate or to evaluate this data.
I'm Michael Chow. I'm on the open source team at Posit. So I work on tools for data analysis in Python. So I think word on the street is that for every piece of Posit gear people have on, they get three minutes of advice that they can get from me. I don't know how great I am as an advice giver. So maybe that's, they're putting me to the test too. But I'm excited to be part of the competition.
My name's Elijah. It's time to check in with some of the creators here at the Posit hackathon. Do you want to be an analyst or something? What's your first impression of the data? I'm part of our team's what we're codenaming Project Phoenix. Top secret. Can't even talk about that. I can see right here. You've Googled what is Python. So it definitely seems like your team's onto a good start. What's your all's team name? Team MMA.
You got to love the charisma. What's your all's team name? Did we come up with a team name? I don't think we did. No name. No name. They seem to be on top of things.
So what do you think about the data so far? Oh, well I guess it does. Some people are doing a lot of talking and not a lot of typing. So I'm just curious to see what they output because we got a lot of people doing a lot of talking. I appreciate the brutality. Back to you, me.
First of all, Harpreet, what is your team name? Positively skewed. What do you think about the data so far? It seems to have rows and columns with values. In them. So I have a job. Unlike all these other people. So that's part of it now.
Shashank, he is working as a team of one. He actually won this hackathon the last time they did it. Shashank, how are things looking for you so far? You know what? I think this hackathon just comes down to ideas. And I think I have a very good idea. Or at least an idea that I'm very happy with. So even if I don't win, I'm going to take this idea home. I'm going to build on it. And I'll be able to use it in my own personal life. We have a lot of talented competition this time. So I'd give myself a 30% chance of winning.
Modesty doesn't sell. We'll come back here later and hope for something else that'll get some clicks.
So we can do a trick planner based on where you want to go from A to B. And then based on the times in which the bikes are used, we can say like, hey, you're going to go at a busy time here. This bike place in your location is going to be horrible. Don't go next block. How does that sound? Interesting.
So this is Elijah Butler back at the deposit. So you think you can analyze data deathmatch. The winner gets the head of their opponents. Bragging rights. They get bragging rights if they win this. There's been a lot of tears shed by everybody in the competition except for you. So how are you managing to keep your composure during this competition? I cried a lot the few days before. So I have nothing left to cry.
I think the one thing we did is we did user interviews of the judges. To understand their preferences for bike shares. And so we're going to deliver to what they need. That seems a little cheaty, but this reporter isn't going to make any judgments.
So right here, we have one of the judges of this elite hackathon. Julia, what do you think of the competition so far? You know, some people have some real problems even getting started, getting things installed, getting things working locally. Classic data science problem, right? Like how am I going to install my dependencies? But it looks like everyone's gotten over that hump. It looks like people are actually working now. And so it's exciting to see how it's coming together. Are there any judges that you would like to talk smack about? All the judges outrank me and our company. So I don't think that would be very wise of me to talk smack about them.
This journalist agrees. You seem to be a person that I could finally get some smack about somebody else in the competition. Is there any smack you'd like to talk? Oh, for sure. We rule and all the rest of them drool. I mean, it's clear we're going to win. And I mean, I don't even know why they're bothering.
We rule and all the rest of them drool. I mean, it's clear we're going to win. And I mean, I don't even know why they're bothering.
So back here reporting with all the hard-hitting journalism, Avery, what do you think about all of this competition so far? Oh, it works. Look how pretty that is. I haven't seen anything that beautiful since my vacation to the Grand Canyon. I think Keith is overdoing something crazy. And I don't think it's going to work out. We'll have to see. I don't think what I'm doing is going to work out. Jack, are you doing good stuff? Sure.
So we're here with another one of our judges. Joe, how do you think everything's going? I really have no idea what to expect. I'm trying not to ask too many explicit questions and just sort of peer over shoulders. And I definitely see people debugging what look like interesting things. It seems like people are mostly getting along really well. And I even think I saw some cross-team cooperation and helping debug something. So everyone's getting along. So there might need to be some artificial insertions of drama into this. While we had the cameras off, this team right here was caught screaming at each other.
Alex, just in one word, how is this competition going for you all so far? It's very stressful. More stressful than I was anticipating. But I'm having a good time. And I have high hopes for whatever we are or are not building.
We're here with another one of the judges. Carlos, what's the craziest thing you've seen out there during this competition? I think folks are being super ambitious and sort of trying to do things that I would be happy if I pulled off in a day of work. And they're trying to do it in two hours. And it's really fun to watch. But I think there might be a little bit of like a last minute drama of like, will this train wreck or it won't? Maybe I'm a softie. But picking the best one of them will be challenging. I think folks are doing a great job. 48 hours in, no food, no sleep, no access to their families. Things are getting nuts.
First of all, I've been worried about you. How are you doing right now? I wish I would learn how to code in high school. So I'd be better and I could win. Maybe we should use our lifeline. I need like shiny help. What is your old strategy going into this? We're working, bro. Bro, we're working. We're getting somewhere. We're getting really close. But I think we are going to have a little bit of strategy where we take some positive paraphernalia, get a little bit more time with the experts.
A lot of the other competition has described you as a total sham and undeserving of anything good that's ever happened to you. How do you respond to that? I respect it.
Shashank's presentation
The first team slash contestant is Shashank. Please come up, Shashank, and present your solution to the judges. Thank you so much, everyone. I'm here to present my idea for an app that I built using Shiny Express. I actually was a big user of the RStudio suite of tools back in my previous, previous job. And these apps in Shiny just looked a lot better than almost anything we can make in Python in the same amount of time.
One of the people I find most inspirational when I'm trying to think about an idea is the creator of Wordle. So the guy that created it, his whole thing was he wanted to create an app that his wife actually used because his wife had no interest in his career as a software engineer. And so when I'm trying to think of an idea, my significant other, she has very little interest in my career as a software engineer. So I'm like, OK, what could I make that she would find very useful? So the other day she was telling me, Shashank, you're not very romantic. I want you to be more romantic. I'm like, OK, what does that mean? And, you know, normally another schmuck would go and he'd be like, OK, I'm going to go try and figure out how to be romantic. And I'm like, no, I'm going to get a computer to do it for me.
And so what I present to you guys is your solution to all your romance problems for the city of Chicago specifically to be expanded to future cities. So what this tool does is this tool, it goes to Google Maps for whatever city you input into it, and it will go ahead and pull a list of 35 locations. And it goes ahead and it picked three random spots inside a radius, provides it to chat GPT or GPT-4, I think. And the GPT-4 will go ahead and give you a romantic date in this case, in the city of Chicago. It'll tell you exactly what to do, why you should go there.
I have a lot of experience with these, we call them city bikes in New York. One of the problems I've always had is that they're all based on stations, right? You sometimes go to a station and there's not a bike there. And so what we do over here is we actually cross-reference the locations I have over here with the nearest three stations. And then we find the station that for a given time of day, in this case, I just picked 12 p.m. for everything, is most likely to actually have a bike available and where you deposit the bike is most likely to have a spot that you can put the bike back. The whole point of this is to provide you with the most, let's say, low-stress way of planning a date, executing a date, so you can focus on what's important. And that is paying attention to your partner and telling them how important they are to you, because a computer took care of all the grunt work.
Thank you, Shashank. Judges, do you have any questions? So you did all this in only, what, an hour and a half or so? But I'm curious, if you had another hour and a half, where else would you see yourself taking this? What I was really looking forward to was having this deployed on a mobile app and just having a QR code over here. And then you guys went and looked at it yourself. I don't want to comment on other people's projects, but I think I could get it deployed faster than most people could.
If you were to start again, are there things you wish you knew right when you started that would make you develop this differently, sort of explore different things? One of the big things that I did is I used Haversine distance. So the distance by how the crow flies for the distance between the different bike stations, which is not necessarily accurate. And in such a small distance in a city, it should be fine. But I would really like to go to the Google Maps API and actually get a path and then use a GeoJSON object to plot it on top of this map. So it looks more like a Google Maps native view. Thank you, Shashank.
Judges, what did you think about Shashank's performance? I feel like that's going to be hard to beat. I mean, that was quite impressive. I feel like the things that were the most impressive were it really did demonstrate quite a bit of creative thinking. I really loved his presentation. He seemed like he had a story to tell. He was confident with it. I was really impressed. And that's a real strong start.
I really loved the way in which he used ChatGPT as sort of just a serendipity generator. The fact that he could correlate that with the data that exists on this reference set and then go search against it and present it in a map means that you get this grounding of the results and you get the narrative next to it to sort of support it. I thought it was really strong. Amazing. Thank you, judges. We'll bring in the next competitors.
Team MMA's presentation
Welcome back. This is Team MMA with Mickey, Monica, and Alex. Good luck. When we looked at this data set, we were trying to think of something that would differentiate us from our competitors who are going to do some boring data analysis, which I mean, come on, guys, a dashboard. Like, we need something better. A user is going to input their address, date, and time. This is when they're going to need the bike in the future. We're going to take the data that we have and forecast into the future a time when they're inputting this information, and we're going to project whether a bike is going to be available at that time at the nearest location.
What it will give as the output or would give as the output is a map with the location of the nearest station that has an available bike. If it does not have an available bike, it would not be part of the output. It would also give the distance to that station based off of your longitudinal current data. And so as you can see, we don't have a map down here. We were this close, and in fact, we could have hard-coded it and we could have deceived you guys into thinking it was working, but we chose not to do that.
The first way we could approach this was a purely, like, a time series machine learning problem. The problem was essentially granularity in terms of seconds and minutes. So then we said, okay, what if we kind of just did essentially like a dummy sort of lookup? So user inputs an address. Using that address, we use a GeoPi package to then get the lat long. We would then also, using a Haverstein distance, we would then get the lat long of all the different stations. And then we essentially say, okay, we then convert it into kilometers. And then we say, give us like locations and then we return it. And so that is the intent of the code. Thank you, Team MMA. Judges, do you have any questions?
How long did that last issue that you were facing trying to get that map to render? How long were you working on that? We actually divvied up some of the tasks so that we can work on things at the same time. I think that lasted about 45 minutes. The difficulty we had was understanding when do we switch from like vanilla Python to like essentially Jupyter, like Python, which is what we're using. And then how do we then make that final translation into shiny Python with like widgets? And then that was actually really where we got stuck was essentially like the input text for the fields and then just returning the values and making sure it would like flow through the rest of like the logic. Thank you so much, Team MMA.
Judges, what did you think? It is, you know, it's such a heartbreak. Such a heartbreak, especially when they felt like they were close. They did not really start with the existing app and kind of minorly change it. They seemed like they were starting from scratch. So that is quite ambitious. And I like to see that. I really like the fact that they went big for the thing that they built, which is if that existed, you could, you know, package it up and ship as like an app for people that would want to use. So more than just a demo of the technology, like that was a real thing that they just said, this is what we're going to try to do.
If all that was taking the better part of three hours, then it was not realistic to expect integration to take about 15 minutes. It's a strong contrast to, you know, some of the other competitors who kind of knew coming in exactly what their strengths were and then like add one new thing to it as opposed to, you know, doing a lot of things that were a little bit of a stretch. Real bummer, though. I was really hoping to see that come together. Amazing. Thank you, judges. We will bring in the next team.
Team Jack's presentation
Team Jack of Jack, Keith and Avery. Welcome. It's now your time to present in front of the judges. Who here has seen bike shares? Who's seen like delivery trucks of bikes being dumped? We can all understand why, right? Because the system doesn't work perfectly. It's not perfect cycles where you have a network of nodes and you have bikes going to and from various nodes. So the problem that we're solving for essentially is that there is an overflow of bikes going toward any one given node. So I grew up in Chicago. I lived in Chicago. And the biggest problem is that everyone either goes to the loop or comes from the loop. And then there's always an issue with bikes being dumped somewhere.
What we're trying to solve for is how do we know at any given time, what is the capacity at this node, at this station? If we can know at any given time, what the capacity is, how it's changing, then we can react and we can send these bikes to and from nodes. So that's the problem that we're solving. The way we chose to look at this was with a geo scatter. I love geo scatters because it lets you see a lot of detail. You could zoom in, zoom out of detail. And so we actually ended up creating is an interactive geo plot here that basically shows you throughout time for all three cities, what the capacity is. So we basically took all the stations and figured out how many bikes can each station hold at any given point in time, on average, how many bikes are available.
And so Keith, if you want to press play, we have from midnight to basically midnight throughout the day. And it's going one step at a time. And these dots are changing based off of that. With the dashboard that we created, we actually have the ability to show you guys the capacity from zero to one. We can show you which ones are dangerously low and need someone to like a truck to be sent over there at any point in the day. And we can also show you how many bikes are available at any of those points throughout the day.
Hey everyone, I'm on the experimental research group at the BikeShare dashboard company. And Avery here is a new dad. He doesn't always have his hands available to look at dashboards of bike availability. We have this new feature we've incorporated into our app. And if I start yelling and being really loud, then we will see shortly that, as you can tell right now, you're really zoomed in and getting proper coordinates. So in two hours, what we were able to do, if we can get this to work, is we can take volume or audio as input and we can use it to control the visual. Ta-da! Thank you for watching.
How did this all come together? Was this like sort of the original plan or was this something that, you know, there's a bunch of just experimenting and then it sort of came together at the last minute? This was our original plan. We always had Keith in that department, that special Keith department. So he was always working on that. And then Jack and I, I really like to see things visually. Like, how do you visually see what 190,000 rows looks like? And so, and I love with animation too, because I can like tell a lot of, a lot of stories in one individual things. But what I realized is there's like five stations in most cities where there's like 50 bikes and the rest of the station is usually in about four bikes. So when you're looking at bike availability, it's always going to be skewed pretty heavily. So learned some pretty cool patterns from the day.
I think what was challenging was trying to balance the stuff that was given versus how I normally approach data analysis. I'm building off the work of others, not inspecting for the first time, while also trying out a new library in terms of visualization. So it's sort of like there was like two to three axes of newness, which I underestimated the challenge that that has, especially given time constraints and given a team working together. Thank you, Team Jack.
Judges, what did you think about Team Jack? They had that problem with the animation with Plotly and they sort of bailed and had it like just load separately. But I liked that they did that and didn't just sort of can the feature, you know, like recognizing that there was something of value created here, even if it was not exactly in the way that they intended it to be. I think it's pretty interesting that their instinct was trying to show an animation in an interactive app, but I thought it was worthwhile for them to engage in that. And sure, it would have been great to see that working. The most actionable sort of like vision of the ones we've seen so far. I really like, again, the fact that sort of folks thinking on their feet and say, OK, we can show what we meant by this on a separate place and so that you can see what we were trying to go. And I really appreciate that. Amazing. Thank you, judges. We'll send in the next team.
Team Positively Skewed's presentation
Team Positively Skewed, it is now your time to impress the judges. There's actually not much to show, but I think it's a fantastic retrospective. Instead of creating a business user focused app, we wanted to create something for an actual customer to use. So when we when we analyze the data, we probably spent a little too much time trying to figure out what some of the columns meant, which is really great in the workplace, but probably not as good for a hackathon. And then we eventually settled on a plan of of creating basically a button that simulates a person.
So we would take the by city. We would take the min max lat long and the min max time and then simulate a person being in one of these cities during that time period of the data, which is about four days. Then we would basically find the closest stations which had an available dock, the top three and the closest with a bike. So for returning and forgetting, probably the biggest challenge would have been doing actual GPS distances versus just have a sign like linear distances. However, the only the only thing we got to was actually calculating those numbers. So we got like random timestamp and lat long generation by city. So if you click it, it continues to generate them. And we just printed it just to show that we did all this stuff. I was working on the front end bit and then Harpreet was working on a lot of these helper functions. That was actually a really effective workflow. Just not enough time.
Judges, do you have any questions? Yeah, I'm curious how much how much time had you planned to give yourself to implement the sort of interactive UI portions of it? We didn't actually try to allocate specific blocks of time to doing the different steps. We tried to do each step essentially as fast as we thought should be done. It's something I'm very used to in what I do and in my background. And I guess like a hackathon requires a different kind of like approach. I think we spent like literally 80% of the time checking on the data and thinking that the 20% would be super easy. But when the time came, it was like, oh, no, that 20% is not enough. Thank you so much, team. It was positively interesting.
Thank you, guys. Judges, what did you think of that team? It was great to see when that reactive thing came back that they like hooked up the button to doing something and that they were able to get that kind of like reactive loop. That was really good to see. This team maybe struggled with time management a little bit. The process is as interesting as the output and seeing how people approach these challenges. And for this group, you can tell like they're seasoned professionals and they know that you need to watch out for pitfalls in the data. It is a shame to sort of see folks having a plan and like all of this code written and then at the end of the day, you hit the deadline and you don't get to see the benefits of that. Avoiding that from happening, I think it's sort of like a good way to think about your future like hackathon participation. Amazing. Thank you, judges. We will send in now the final group.
Null Consulting's presentation
We welcome Null Consulting. It's now your time to impress the judges. We're Null Consulting. We're working for bikesarefun.com given the data we have. This is their lovely logo. It's very professional. This is made with Quarto in spirit. Bike sharing is hard. You can see all these horror stories, but we think this is an opportunity. The problem, low-use areas isn't a sign of poor performance. Instead, there are opportunities to increase revenue. But how? We did extensive research with Userbase. It was Julie. We found out the following. Users often use the bikes not for commuting, but for leisure activities. This was especially true when they're traveling. Bike services are seen as an amazing way to explore the city.
So the solution, rather than ignoring low-use areas, we instead need to encourage users why it's worth visiting these areas. So welcome to Bikes Are Fun City Guides. By tying restaurant entertainment data with low-use stations, we can unlock a completely different market and grow your business. Well, we used a model to essentially find the low-use areas. We used interquartile range to essentially figure out which ones we could focus on. In each of these sections, there's about, I don't know, somewhere 30 possible areas that we noticed. And we decided to make sure they were mapped out. And basically, the next phase of this project is to start scraping the internet for different areas, different lats and longitudes, that have interesting spaces that we could then start pushing to your guys' app to encourage your users to go to those spaces that could then increase travel and use of these bikes.
I kind of like to think about it like a Michelin guide, but for bikes. We weren't able to do that last step, but we started scraping data from Michelin and TripAdvisor for top points of interest, restaurants, to fit in with this theme of routing people to stations near restaurants and using that restaurant data to just guide people and help them explore the city while also using bikes for fun. Any questions?
I'm thinking more about you trying to use a model-based or a stats-based way to find the things you were interested in. How was the process of thinking about that within the reactive framework? I had originally separated the stations and bikes data separately. And then there was an issue where it was trying to calculate something that I was expecting from one of the other calculations that was basically going to run all the time. And so then I was like, oh, this is not there yet because they're constantly rerunning. So that was one of those things that I was like, what's happening? And then I kind of figured that out.
If you were to start over, where would you do things differently and how? What are the lessons you've learned? Exploring the data, that's where we start it. And you can get lost in that, especially for a hackathon because it's so fun. Good data people should start with data exploration, but within the confines of a hackathon, maybe two of us could have started with data exploration and one of us could have started with sample Quarto dashboards and starting to understand that a little bit better. So we kind of saved that part for the end because we first wanted to know what data we were looking at. And that had us rushing towards the end. We could have split the work a little bit better and had one person focus on Quarto. Thank you so much, Null Consulting. You may now exit.
Judges' deliberation
Judges, what did you think of Team Null Consulting? I really liked their, I guess, the sort of product side. The humor was great. The user research was hilarious. I mean, not just hilarious, but also like that is what you would do. You know, if you were looking into this, I thought it was really fun from that perspective. I'm probably going to give this team my highest presentation style points because I feel like the way that they presented their work was funny, fun, but I think also really demonstrates an understanding of what it is like to build data products with people.
This is the dashboard that you will see in a company somewhere, right? Somewhere someone will make the dashboard that has a functional basic visualization that has an actual model behind it that from there you can guess what is going to happen next, right? So is it station by station? Is it region? How do you like aggregate it? What do you do from there? Indicative of what you do like in a real world.
So judges, what are your final thoughts? I have to admit, I've got a real soft spot for Team Jack. You know, I love in hackathons when someone takes the moonshot. The problem we solved was real. I thought our vision was genuine. And that microphone thing, I mean, we were surprised that they even attempted something like that. If I start yelling and being really loud. Ultimately, like it didn't quite work in the demo, but I do believe them that it was working an hour ago. And I mean, that would have been easily the most impressive single feature that we saw all day. The fact that it didn't work was a minor technical glitch.
I do find the work that Null Consulting did so compelling. I mean, part of it is the humor and the presentation. But what underlies that, like what made it land in terms of the humor, it was a substance behind it. The competition was fierce, but you know, we trained for this every single day and came down to execute. We did it. I really liked that Shashank pulled sort of data from ChatGPT and then cross-searched it with like the Geo database and then showed the results. I presented something that I think is genuinely useful. And that's all I can ask for myself. I thought that was a really interesting demo. And then to have it end with like, and you reload it and you get a new date. I thought it was a sort of nice bit of showmanship. So I was a fan of that.
Presentation went well. Unfortunately, we had an unfinished app. Our presentation was awesome. We worked extremely hard and I'm really proud of where we got. In the end, we were really close.
Winner announcement
Judges, do you have your final decision? I think we know. Yeah, I think we know. I think we do. We are going to announce the winner of the So You Think You Can Analyze competition. We learned a lot. There was a lot of struggle. There were some tears, but there was also a lot of smiles and a lot of fun. We will leave it to the judges to tell you first who the runner-up was and then who the final winner was.
The team that we identified as the runner-up is a team who we were impressed with their humor, their practical thinking. Honestly, how well we think what they did would have landed in a real business environment. Not doing everything they planned, but really showing iterative, complete work. Our runner-up is Null Consulting. Where's the rest of your team? I did the least work on the team, but because my team members are not here, I'll take all the credit. So thank you.
And the winner's app. And by the way, it was pretty close. Pretty close. We had to check the math. The winner's app had a little bit of everything. It had humor, it had grace, it had a clever story. There was a little bit of love. And this team made it look relatively easy. Operating under some serious constraints and disadvantages, and yet pulled off a really graceful app and a graceful presentation. The winner is Team Shashank.
The winner's app had a little bit of everything. It had humor, it had grace, it had a clever story. There was a little bit of love. And this team made it look relatively easy. Operating under some serious constraints and disadvantages, and yet pulled off a really graceful app and a graceful presentation.
I had a great time doing this at the end of the day. I have, I don't know if anyone watches football here, but Bill Belichick, right? This is, you know, this saying, trust the process, right? I think what I've learned is that, make a good process to do things and don't worry too much about the results. And if you keep doing it, you keep iterating on it, then you'll eventually, you'll get somewhere where at least you're happy with the work you put in, you know? My girlfriend does not care about software engineering, but she did tell me I'm not romantic enough. And so I'm like, okay, how can I offload the, how can I offload the work of being romantic to a computer? And I think I successfully did it. I showed it to her and she's like, wait, this is actually pretty good. So, you know, there you go.
You're a former Iron Analyst champion. Now you're a So You Think You Can Analyze champion. Will anyone ever dethrone Shashank? All right, thanks everyone. That was amazing. Appreciate it.
Oh my gosh, he's still here. Go home. It's working now. Are you kidding me?