How to use ChatGPT Code Interpreter for Business Data Analysis

say goodbye to excel because chat gpt's code interpreter is here the latest model from open AI ensures that the language model can easily understand python code analyze data from Excel sheets CSV files and other popular formats and reply to you in natural language to get started you need to have a paid account for chargeability and then you will be able to access the code interpreter model by going to your account settings and in the beta features you can select plugins and code interpreter turn on the code interpreter model so that you can access this model and upload your files and start analyzing data easily when you activate code interpreter you will have to go to gpd4 and then select the code interpreter model when you activate this model you will see the new plus icon here which allows you to upload files and when you click on it you can head over to wherever your Excel file is so when you open this up now you will have this gpt4 model and you will have the code interpreter option that you can select in order to start uploading files and analyzing the data once I click on it I'll find that there's a new plus icon over here that allows me to upload files and when I click on it I have the ability to easily upload files now so in the Excel file that I am uploading here are various uh there are various spreadsheets and each of them contains data specific to that spreadsheet and the data also works across spreadsheets because the balance sheet and profit and loss statement are very much related then you have the child balance which is used to create these financial statements so we'll be able to see how it can analyze data from a single spreadsheet and also from multiple spreadsheets for which you practically would use various for which you will practically would use various other tools such as Visual Basic and then your use may be of a scripts and other very complex things and Excel specially power pivot and power query and you don't have to learn all of that if code interpreter actually succeeds here and analyzing all of this data for you now it has uploaded the data and I am going to ask it to analyze the data for me but first it has to clean up the data now part of working with data as you may know or anyone who has ever worked with data is cleaning the data up because it can take a lot of time to clean up your data to properly format it to properly structure it so that when you are analyzing it you don't actually fall into any pitfalls or encounter errors so I am going to ask it to format the data and then show it to me in a workbook in the form of a table using markdown syntax which actually allows the program to show me the table format okay so let's see how it works the good thing is that it will actually show you how it is cleaning up the data and the thinking process that it goes through okay so it has access the spreadsheet program and it has listed all of the spreadsheets that are in the workbook so that's the first step and it has done it successfully and it is giving us its own thinking process okay now it is giving us the data from the child balance but that should not be that's actually not correct because there are a lot of other data in it foreign specify that it has not given me the entire data set so it'll again go through it and again format the data and use the and replace the nand values with a zero as you can see it has started formatting the data and I'll go and look at the data and I can see that the values actually match so that's a good start so I am giving it further instructions to remove the data that is not needed because it has taken the date and the empty columns to also have a value in the trial balance and that is not something that I want and this clearly shows that simply I want so it clearly shows that it can get things wrong and I'll have to again and again Ask it to clear up the data and format it properly so it is not something that you can you know trust blindly because these models are prone to something called hallucinations so they invent the data automatically they can conjure up new facts and that really makes it very important for you as a user to verify the data again and again okay so we now have the data properly formatted and I am going to ask it to create a profit and loss statement and a balance sheet on the basis of this trial balance let me just zoom in a little so that you can see more clearly and it is going to create a balance sheet and a profit and loss statement for me and we'll cross check it with the one that I extracted from quickbox online [Music] okay so it thinks that it has not actually identified all of the topics properly and categorized them so it is very important for it to categorize the transactions so that it knows in which uh what accounts belong together what accounts are expense accounts what are Income accounts and then it'll again categorize them and start creating the financial reports for us okay so it is still working I am going to see if the profit and loss statement if the balance sheet is actually correct or not and I'll open the balance sheet up in the Excel file and here I'm going to match it so that I can easily see if it is correct or not okay so like this it has failed in a big way here because cost of goods sold definitely does not belong in the balance sheet it is a profit and loss it is an income statement account it is a temporary account and its balance has to be transferred to the retained earnings account or to the owner Security account before the end of the year so again as you can see we have to format the data we have to give it very clear instructions okay so I have asked it to reject the data that it created and I'm going to get it to format the balance sheet and the profit and loss statement I provided through the Excel sheet which I downloaded from with box online so that we have some real world data to work with okay it is giving us all of the values and the problem that we are coming across again is that the data is not correctly formatted so I'll have to again try with the CSV file that is properly formatted so that I can get straight into the analysis side of things [Music] and to be fair this whole file that I extracted from good books online as a complete mess there are indentations there are okay so we have the balance sheet and the profit and loss statement okay now I'm going to start asking that questions so that I can analyze the data so if you don't know current ratios simply is a liquidity ratio that tells you if a business has enough current assets to pay off its current liabilities that is made Financial Obligations in the immediate future [Music] so it gives me a brief explanation of what the so on the basis of very specific examples it is again going and reformatting the whole data okay so it has calculated the current ratio to be 1.62 and it shows that it is really good especially for a design agency we know that the quick ratio which is specifically more geared towards merchandising and Manufacturing businesses that hold on to inventory is uh is not that much different from the particular current ratio when it comes to service based businesses I am going to calculate the same thing in Excel so I'm going to first name the ranges so I have total current assets over here and I am going to define the name and I also have total current liabilities who is Ohio current ratio and it is current assets by current liabilities and jacobd was able to correctly calculate the state of order okay now I want it to calculate some other important financial ratios on the basis of the balance sheet and the profit and loss statement and then we are going to go back to our spreadsheet and calculate them manually to see if it is able to actually calculate the values without much of our intervention okay so I'm getting two other uh financial ratios from the balance sheet and it is going to first tell us what these are why they are used and what is their relevance how to calculate them and then it is going to work with the data in order to calculate the whole amount okay so it has calculated the tattoo equity ratio as a negative number which means that the debt is too high for the company okay so I'm going to verify this data using the Excel spreadsheet and I'll calculate the tech to asset manually and to do that I can go and select the total liabilities and then divide it by total assets and their Tech to equity Tech 2 asset is 1.32 and debt to equity is obviously we have total Tech 2 asset ratio as 1.33 so it has rounded it up and for gross profit ratio we'll calculate it later but that to equity ratio isn't correct foreign okay I'm asking it to verify the data once again and let's see what it gives us okay so I am getting it wrong that was really messed up from my side and I was able to verify my incorrect assumption that I was doing things correctly and I am going to divide it by total equity and here I have okay so it is right but now I'm going to check other things such as the operating profit ratio the gross profit ratio and the net profit ratio and for that I am going to head back to Excel and open the profit and loss statement and here I'm going to name the total income as and I am going to give it a name so that I can use it again and again then I have the operating profit the gross profit and where do I have the cross profit so I'll name this as well then I have the operating profit and finally the net profit okay so I have named the ranges I am going to enter the name and these are all ratios foreign I was given so the gross profit ratio is 0.96 and that's correct and then the operating profit ratio is 0.45 which is also correct because it has rounded it up and the net profit ratio is 0.16 and that 2 is correct so it can perform calculations correctly and very appropriately very quickly so it makes financial data analysis a lot easier for people like me and you okay as you can see here I was able to verify if the data presented to me was right or wrong because this time I made a mistake personally and I cross checked it with the model and it was able to verify the calculation and you will have to do that again and again because it can get things wrong and you can also get things wrong after all we are both intelligence maybe one is natural and one is artificial okay so what I want to do now is get some I want to get some more data and for that I mean go to channel and this is a website that hosts a lot of data sets that you can download and use for practice exercises to maybe analyze and I'm going to use data from here to see if we can get really good data analysis from this program okay so let's just search for a CSV and not on the basis of hotness but on most quotes and we have 19 open research Netflix movies and TV shows so maybe I'll go with the Netflix movies and TV shows because that's what I am into more of and I'll download the data set I'll of course have to login with my account so that I can continue with the download and the good thing here is that the CSV file is properly formatted so that the program will be able to easily understand what the data is how it functions it is not something like what we had with QuickBooks Online the data is well structured here and again click on download wait for it to download and I'm done okay now I'll go back to chat.openai.com so that I can you know upload the file and see again g34 and I want code interpreter to be enabled so that I can upload the file and okay so here's my file and okay so it is giving me the column names for each of the headers and that way I'll be able to quickly know about the directors the movies that they have made for Netflix and the countries that have been targeted so I have all of these things here so we know the structure of the data and we are ready to proceed foreign okay so we are going to get into some data visualization and so the so from our last chat we know that it wasn't installing matplotlib and now it says that it has already has it installed so there is a lack of consistency of course and I'll go ahead and I'll say okay so it is working now and it will go through our data and it will try to find out and condense that data so that it can then give us a really good graph yeah so this is a really nice craft it has the appropriate X and Y axises and it defines the names of each of the directors and also color codes them so that they are different and the chart looks really good movies here and I'll go again with code interpreter I'll download a new data set from kaggle and this time not Netflix movies and TV shows because I can clearly see that the model is not that comfortable with creating data and analyzing it for this type of thing okay I want CSV file so I'll filter them out and I'll sort it on the basis of most votes so that I can easily find the best data sets that we can use for now okay so I'll go with code reads and I'll try to find out what kind of data I can get here and I have the book ID the title of the book the author the average rating the number of the book the ispn number that is a unique identifier then we have the rating count and this looks like a really good data to analyze I download this data I'll save it and I'll also first unzip it extract all adapt data now I'll go back to chat GPT and I'm going to click on the plus icon over here to add my file in books I'll select the CSV file okay so now I have given it a prompt to give me the top 10 books on the basis of the ratings here and let's see if it is able to give me that and the thing here is that I haven't asked it to rate on the basis of number of ports that it has gotten so these are some of the books that may have less number of ratings but they may have good ratings like maybe only a single person voted and gave it a rating of 5 and that would again become a top 10 Stock in terms of rating foreign okay so now instead of ratings and that is the average ratings I wanted to give me the data filter it out and sort it on the basis of the vote count okay so now we are getting some titles that we are very familiar with especially Twilight The Hobbit Angels and Demons The Catcher in The Rack so now we can see that we also need to ask at the right questions to find out the most popular books Maybe we have to see which actor that we have to use to filter and sort the data [Applause] okay now I wanted to create a graph and it has to sort of automatically figure out what type of graph that it should create and then create a graph accordingly so it has created a bar graph and that is the most appropriate graph here to compare data that is in different rows okay now I want the top 10 authors on the basis of the number of books that they have written and it is giving me that data thank you and I I believe that I want a very specific chart for this so I wanted to create a I wanted to create a bubble chart editing because that is also a very good example of the new things that this model can do and I upload a basic image foreign okay so I think it is going to be able to perform this action because there are python libraries that enable this kind of functionality and let's see if this model is able to do this right now and this is really exciting because it opens up a whole world of new kind of software in which you will be able to edit videos edit pictures just by simple text prompts so learning the complexity of different software won't be an issue for most of us then [Applause] okay so I'm going to have it reduced the size of the image and let's see if it can function with it properly because right now it gave an error that it is out of memory so how it works is that this whole operation Works in a sandboxed environment think of it like a smaller computer inside a larger computer and the amount of resources are very constrained in that smaller environment so maybe that is why it does not have enough memory to perform the action and maybe as this feature rolls out as more people start using it uh maybe the company expands the ability to create to work with more to work with higher resolution files and let's see what we get here foreign is really animated then this will just blow my mind right now and the file size is quite large because it is 82.8 MB so I guess even so before maybe it reduced the size maybe it reached over 100 MB or maybe even more and I launch it and I mean this just blows my mind like it is really animated the whole image it is a very slow Zoom effect that is really kind of good in this way and I'm also going to use another function on it which is to apply filters to the image so right now we have seen that it can animate images it can resize images but what about simply applying basic filters a lot of us do that in various software such as Photoshop or we use canva or other online platforms but let's say foreign there are a lot of use cases for this especially if you are running a small business you have to start using this tool in order to maximize your gains in order to analyze your marketing data your Finance data and to maybe even work with images and sort of explore what other use cases you can come up with and if you have tried it out let me know your experience in the comments down below and maybe you can have a great discussion about it and maybe other users can learn something new from it and if you like this video then subscribe to our more awesome content

Channel: Dancing Numbers

Views: 289

Keywords: accounting, how to use chatgpt code interpreter, openai, openai chatgpt, gpt 4, code interpreter, python, data analysis, artificial intelligence, data visualization, data cleanup, microsoft excel, google sheets, spreadsheet, quickbooks, small business, business analytics, natural language processing, excel alternative

Id: hbLJA06bTmg

Length: 53min 21sec (3201 seconds)

Published: Thu Jul 13 2023