ChatGPT Code Interpreter - A New Era of Data Science Begins

open AI just released code interpreter for chat GPT plus user so hopefully you have access to if you are a member of that so I thought in this video we are going to take a look at how we can use this so we're going to test out some Tesla stock price data sets we're gonna test out some sleep analysis data set I found on kaggle I'm gonna test out some images from mid-journey can it read those I don't know we're just gonna have to explore that my content calendar I want to try if I can see some video ideas from all my titles from my YouTube videos I want to test out some python code can it explain that can I can create something visual from that okay so a lot of things we have to explore here so like I haven't tried this before so this is gonna be my first impressions so yeah let's just give it a spin let's just start by showing you how you can activate this if you are a plus User it's quite simple what you have to do is just go down here click on settings go to Beta features and just click on code interpreter here and just select gpt4 here and code interpreter and that's basically all you have to do you can see you have this little plus sign down here and here is where you can upload files so we are going to start off with the sleep and lifestyle data set I just downloaded here from kaggle so this is a great place to find data sets if you want to test this out so this data set is basically comprehensive sleep metrics lifestyle factors cardiovascular health so it's basically kind of a diverse data sets so we're going to ask this and see what kind of information we can use code interpreter to get out of this data set so we're just going to start by asking simply explain what this data set is let's just do that so we can see we have it uploaded here so just click submit and now you can see it's working so basically what it's doing here it's creating a python code and it looks all at this data set using something called pandas as PD so here you can see this data set appears to be Health sleep and light style habits for barriers individuals so we have gender age occupation sleep duration quality of sleep physical activity level so basically it reads everything we need to know about this data set but this is fine but what can we do to get something out of this so I just want to type in what trends are this data set showing us let's try that remember this is new for me too wow this is just great look at this so you can see numerical variables so the average age is 42 years old the average partition sleep for about 7.13 hours quality is a score of 731 so we get this really nice details here the average heart rate is about 70 bits per minute with a range and we have daily steps all of this but I want to kind of see this in a visual way so let's ask just ask it like so let's just ask it show four interesting ways to display the data visually oh wow that was quick so here you can see we have distribution of H slit duration by gender daily step versus physical activity level count of sleep disorders okay so let's take a look at the the visual representations here so we have age on this axis we have count here so you can say yeah about 42 3 something here it's the most count yep sleep duration by gender so uh females so baby this is like the average and here we have the distributions I'm not a data scientist but I found this very interesting and here you can see we have some kind of plots so we have daily steps on this axis we have physical activity level here so this goes up to like 90. and like 10 000 90 yeah I guess that count of sleep order so non sleep apnea insomnia okay so yeah I guess we can get a lot of let's ask it something a bit strange let's ask uh give me a sec so let's just test this can you make a 300 word presentation of the data set for my meeting include visual graphs let's just try that and see what happens okay this looks great I guess you kind of missed on the the images though it doesn't come up but this could save you a lot of time right so an overview of sleep health and life cycle lifestyle data set today we are exploring the data set really fascinating insights to sleep habits Health parameters we have all the data at 374 individuals yeah we have the eight range distribution of H sleep duration so is there any conclusion yeah this data set provides valuable insights to interplay within lifestyle habits and health parameters our next step is a bit deeper into this data investigating potential correlations on Trend that can help us understand and promote optimal sleep Health thank you this could save a lot of time if you have a lot of data sets you have to like present in a meeting you can just put it in here write me something like this and just spit it out I guess very interesting let's change it up a bit and I want to upload a python code and ask a bit about that okay so we just start a new window we click on upload file I'm gonna go to my python codes here let's just find I think I made this yesterday let's open that and explain this python code to me let's try that yes perfect so I love this so you can see we have uh this script appears to be a tool for extracting and processing terrain.trending technology news data from YouTube and the web yes that exactly what it is so we have Library Imports filer price functions opening a key transcribing audio YouTube video extraction everything yeah perfect so can we do something visual with this let me try something let's just try can you create any visual representation of the Python script okay so I guess it gave it a best try so load airplane key the fine YouTube channel yeah this is basically what's gonna happen but it didn't get any like good visual representation but I guess it's something so let's try something else okay so I just went over to Mid journey I created this image here so this is basically just New York and you can see we have a woman here looking over I guess this Manhattan so let's try to upload this to the code interpreter to see if it can recognize this okay let's just click upload pick image one okay what can you tell me about this image okay it seems that I can't uh perceive the image you uploaded because my current capabilities okay so there's no visual recognition or interpretation I can perform various analyzers on the image size color distribution and other method metadata just tell me all you know about the image let's just see that okay so we just have size mode and dominant colors so we don't have any like multi-modal stuff yet I wasn't expecting that to be honest but so no computer vision I guess but that might come in the future and that can be very interesting okay that's fine let's move on to something else I wanted to try so let's pick code interpreter I have my content calendar here on notion so I exported this file and I want to see if I can get something useful out of this so let me just go here upload file uh pick this content calendar okay and just ask explain the data set to me okay this data set appears to be content calendar okay yeah Focus about AI so we have name a sign yeah okay deadline link media interesting and let's show visually the distribution of the most used keyword in the titles let's try that oh wow this was cool right so we get these I've seen this so you can see most of my title is chatgpt GPT AI prompt engineering wow I love this this was very cool so let's go looking at all the data in this data set can you make some correlations and brainstorm five more video ideas based on the titles okay so we did actually get some ideas so we got the role of AI in interior design a mid-journey exploration from ID to reality the life cycle of an AI agent yeah I guess we got something uh I probably wouldn't wouldn't push this maybe mastering prompt engineer with AI agent maybe but at least it's very interesting I think I want to finish off with one more data set and try to dive deeper into some more visual representations because I think that's very interesting so I'm back over here on kaggle you can make an account here and they got some trending data set so let me just uh should we grab the Tesla stock yeah let's grab the Tesla stock five years historical data so just click on that and we can download this okay so let me just unzip that head back to church apt and try to make some cool visual uh AIDS from this data set okay code interpreter upload let's pick the training data set we know this is stock prices right so we think we're just going to go straight to some visual representations here so let's just go create a visual representation of the stock price per month over the last three years can this work okay so it's going to use date and close stock price it's going to use a line chart yes this is so good right look at this so we can't really hover it but you can see the price here is around two to three hundred we have the spike here in like late 2021 it was like 400 or something and you can see like lately now we have a new spike in Tesla so that's very interesting I just want to ask if there's some basic simple questions so let's do that let's just do simple what can you tell me about the Tesla stock in the last three years okay so basically I have the trend uh seems to be an upward one indicating the stock price has generally increased over the trade time period okay that's right the note it was the volatility in stock performance the prices expected a number of significant risers unfolds during this period yeah recent performance uh seems a bit downward Trend but it's yeah so it says if you want more in-depth analysis of a specific ass like you're interested in like correlation with the market events please let me know so let's try show me visually how 10 perfect trades in a Tesla stock in the last three years would have looked like okay so this is quite cool so you can see here is the visualization of the Tesla stock price with 10 perfect trades Market the green represents buy actions and the red represents sell actions so you can say here we should have bought we should have bought we should have sold we should have sold bought sold both sold yeah and that is basically it one two three four five six seven eight that's not ten but at least it shows like kind of Simply how we can borrow and sell High I guess uh yeah I guess there's a lot of stuff we can do with this uh I just wanted to Showcase a few things now and yeah very excited for code interpreter let me know are you gonna try this out I think we are just barely touching what you can do with this so I'm gonna make a lot of videos about this I'm going to test different stuff so very interesting hope you all have access get to try this out and Yeah have a great day and I'll see you again soon

Channel: All About AI

Views: 37,337

Keywords: chatgpt code interpreter plugin, chatgpt, chatgpt code interpreter, chatgpt data analysis, data scientist, chatgpt data science, openai code interpreter, how to use chatgpt code interpreter, how to use code interpreter, how to use code interpreter in chatgpt

Id: PBZVCjkTGrY

Length: 12min 33sec (753 seconds)

Published: Thu Jul 06 2023