Getting Started with Data Analysis Using ChatGPT

hi I'm Dave and welcome to the notable chat GPD plugin as we're just getting started here let's take a moment to learn about some of the capabilities of the plugin so we can get the most out of the notebook experience now hopefully you've had success creating your first notebook but if not watch me as I walk through how to create a very simple notebook that simply prints out welcome to the notable chat GPD plugin so as a quick refresher you need to be using the gpt4 model with plugins enabled and have installed the notable plugin so I'll go down to the bottom here I'm gonna give it the prompt that simply says create a notebook that prints out welcome to the notable chat GPD plugin and when you're interacting with plugins you will see these little squares pop up here these little rectangles it's green when it's currently interacting with a plug-in and it turns gray when the interaction is finished now depending on the complexity of your prompt you may see multiple of these rectangles pop up as it's interacting with the plugin it's best to just wait until chat TPT responds with the text with the result of that interaction so let's just give it a second here and let's see what it does we can see here it says that it's created a notebook called the welcome notebook and it added a cell that printed out welcome to the notable chat GPD plugin so there's our first notebook that was created using the plugin now let's click a link to view and interact with that notebook because again when you're creating a notebook with the plugin it is creating an artifact inside of notable that allows you to view and edit and modify that notebook whenever you want so let's click here to say let's view the notebook it's going to open up our notable account and there is that notebook and sure enough it's a notebook that's called Welcome notebook and it prints out welcome to the notable chat GPD plugin at any point we can interact with this notebook and we can even update and modify the notebook so if I want to change that to say welcome to the notebook plugin we are excited that you are here I can do that and I can execute that cell and now you see the output is updated based on what I just modified in that notebook all right now that we've created a simple notebook let's go back and create a new notebook that actually does some real data analysis so let's go back into chat gbt I'm going to give it this prompt here and ask it I want to perform some Eda or exploratory data analysis on the Denver Public Library to see which library locations had the most visitors over the past year I gave it a link to where the CSV was located for that data but I'm going to show some examples later in this video about how you can upload your own data sets or even connect your external database if you have a data warehouse you want to connect to I also add a little bit to the prompt there that says when I'm authoring the notebook always include markdown cells to explain each code cell that just makes the notebook a little easier to read and understand what's happening so such hbt work here for a little bit and then we'll look at the results that it creates and now that it's finished with this analysis we can go back and look to see all the steps that it took one of the first things that it did is it downloaded the data from the URL that I gave it and then printed out the first five rows to help just show an example what the data looks like it did some summary statistics and then to answer the question that I asked specifically It produced a visualization to show which Library location had the most visitors in this past year we can actually see that visualization right here within the chat GPT interface also click the link to the notebook and go and view that with a notable and now we see all the content that was created including the markdown cells that I asked it to create to help me better understand what's happening in the notebook as well as the code cells and the results of executing that code cell so there's the first five rows of the data that was brought back here's some quick summary stats of all the columns and here's that visualization that was created with matplotlib and just like we saw before this is fully interactive so if I want to make any updates to this inside the notebook interface itself I can do that now I want to take a quick moment to talk about where these notebooks live inside notable because it's going to be important to understand as we continue to work with the notable chat GPD plugin notebooks live inside projects and projects are simply a way to organize your work to have notebook files data files or any other assets you want all in one location as part of your analysis so in this case here we're looking at this project which is called Denver libraries Eda and it only has one file inside of it which is The Notebook Denver libraries Eda well how is that possible well at the start of this when I prompted chat GPT to do this analysis it actually created a new project for me and I said all right I'm going to create this project and I'm now going to create a notebook inside that project even though I didn't necessarily ask it to do that it thought it was being helpful to try to organize the analysis to put it all in one project but going back to notable here let's look to see what other projects I have access to so this was the one that was just recently created by chat GPT I can go over here at the top here to switch between my projects you see I have one other project called my first project and that is where that first notebook that was created by that prompt that simply said welcome to the notable chat GPD plugin it created the welcome notebook there now additionally in this first project my first project everyone will see the same notebook that's called what can you do in a notable notebook that contains some helpful documentation for getting up and running but one of the things to know about projects is that chat GPT has the concept of What's called the default project that is when you ask it to create a notebook without any additional context of where it should be created that's by default where it's going to be created so I can go back to chat gbt and I can even ask it what is my default project I gotta spell it right and it should respond with my default project which I'm going to bet is my first project and there we go we see that chat EBT knows that default product is my first project that's why when I asked it to create a notebook without any additional context of where it should be created it created it inside of that project at any point though I can always update and change that so let's say I liked my new Denver Eda Library note project and I want that to be my default project simply just copy and paste the link to the project go over to check gbt and said please make this my default project now for any additional prompts going forward where there's no context that chatgpt has about where would be the best place to create that notebook it's going to default to creating it inside that project which is now updated to be the Denver library Eda project let's talk a little bit more about what we can do with projects inside notable so I'll click a link to the project here and again as we saw before this was a project that was just created for this one notebook so it only has one notebook inside of it but I do have the option to upload any additional files that I want to this project to help me with my further analysis so I'm going to go ahead I'm going to upload some files to my computer here let's just pick a few of these CSV files and upload them and now these files are going to live inside that project and chat GPT will have access to them if I want to do later prompts that reference any of these files you can prove that by going back to the prompt here now saying what files do I have access to that's going to go out do a request to notable to look inside that project and respond back with the same listing of files that you just saw me upload into the project right there so there's the one notebook that I added and now we should have these four data files coming after that so let's say I wanted to create a new notebook that accesses one of these files so let's say yes let's create a new notebook to analyze the NYC squirrel CSV data set and create a map showing the location of each of these squirrels right I can simply reference these files by their file name and check TPT will create a notebook again inside this project that rep that accesses that file and does the analysis that I want it to do which in this case is just creating a simple map and there we go it finished with its analysis it gave me a link to the notebook which I can go ahead and View and just like I asked it accessed the file that I'd uploaded locally to my project called NYC squirrels you can see the full listing of files in the project in the left sidebar there and it created a map that showed the location of the mall and clearly this data set only contained locations of squirrels inside Central Park within New York City all right now so far we've seen examples about how to use the notable check GPT plugin to analyze data that was hosted on the web as well as how to use the plugin to access and analyze data that was uploaded locally to a notable project but let's now talk about how we can use the plugin to access data that was contained in an external data connection or data warehouse so the first thing we need to do here is set up the connection to that external database or data warehouse and we can do that within the notebook page over on the left within the data connection sidebar so you click on the little database icon here and now I already have a data connection that was set up here but I'll quickly walk you through the steps to create your own one click plus create data connection and pick from the variety of different types that notable currently supports now if something that you're using is not listed here please contact us we're more than happy to add support for that as well but if what you're using is here already you simply click the database type that you want and then go in and enter all the information for that particular database type now again I've already set one up so I'm not going to walk through these steps I'll show you what I already have so I called my data connection just simply my data connection and over on the data connection sidebar here we can actually start exploring the scheme of it so you can see here I have this schema for Lake Summit has a couple different tables in it including this weather table now when I use chat EBT I can prompt it to do analysis based on information contained in these data connections so for example here I can start a new prompt that says let's do some analysis of this Lake summer weather data that I've set up in this connection called my data connection and just like I referenced the data files before that were uploaded to the project I can simply reference the name of these data Connections in the prompts to chat GPT and I simply said let's create some charts to understand the changes in the weather over the past year let's kick that off right now that's finished we can go and review the conversation here to see the various visualizations and data that was returned to chat EBT or you can just click the link and view it with a notable again one of the things that made this experience special is that chat gbt was creating content for SQL cells that were querying out to my external data connection so again I had that data connection called my data connection you can see it would automatically create a SQL cell use that data connection and query the data to return as part of the analysis and for some of that data it returned it via SQL and then later used python to create visualizations so we saw this average monthly temperature over the past year we saw additional things about number of rainfall and I think humidity here as well right all created based on that prompt using SQL cells and external data connection in order to get access to its data all right to start to wrap things up here I want to talk about some of the powerful things that you can do inside a notable once you have the data returned in the notebook that was created from chat GPT and for this I want to go back to the New York City scroll data example so one of the common things that check CPT will do is simply ask for a head of a data frame which are the first five rows you can see it here for instance a nice little summary within the chat GPT conversation that just gives you a sample of what the data is but inside of notable we built a lot of powerful tools for doing analysis of a data frame as well as creating data visualizations on top of that and so one of the things you may want to do is just remove that call for DOT head and have it returned the entire data frame and because we do that we now get the entire data frame returned we can see this 3000 rows of this data I can simply scan down through it I can scan across at all different columns I see the distribution at the top I see what data types there are I can change these at any point I can rename the columns I can reorder The Columns I can do a lot of you know quick and easy analysis on the data that's the data frame to include automatically brushing and creating easy filters on top of this but one of the really powerful things that we've done is created an easy no code data visualization capability that sits on top of these data frames called Data Explorer or decks and that allows you to go through and either pick a particular data visualization type that you want and we support over 40 different data visualization types or simply explore with data prism where it will suggest data visualization types based on the size and shape of your data and just like we saw before we know there's a geographic coordinates for this squirrel data and so mapping is a great data visualization type a tile map scrape it data visualization type we may want to explore and at any point we can look through the different data visualization types that data present bought back and we can click on and explore a particular type changing the settings however we want to kind of customize this data visualization to exactly suit our use case all right the last thing that I want to talk about is the notable Pro Plan and some of the powerful features that it provides notable Pro Plan provides access to larger Hardware machines to include large CPUs as well as even gpus as well as additional concurrent kernels and a longer kernel session timeout let's talk about how that may impact your ability to use the notable chat tpd plugin so if I go back to chat TPT here and say create a notebook to analyze the 2015.csv data file one of the things that's going to try to do is to create that notebook like you saw before it'll access the file for my project files and I'll start that kernel up but it's going to run into a limit that I've already hit the max number of concurrent kernels I have created because I have three running notebooks already which is the max limit to the free tier so you see here in a second it's gonna it's gonna say it ran into that and just like I expected it ran into the limit of number of concurrent kernels that can be running in the free tier which is three now it asked me if I wanted to shut down any of those running kernels or upgrade to the Pro Plan a to the Pro Plan would increase that limit up to 10 so when you're working in chat TBT and you're quickly creating new notebooks and doing new analysis you're not limited by the small number of concurrent kernels you can have running another big area you can benefit from on the pro tier is access to Hardware sizes so if you have large data sets and you need to access larger Hardware sizes but either CPU or GPU you can get unlimited use of these larger Hardware sizes by upgrading to the pro tier right I hope that was a helpful introduction to some of the powerful things the notable chat TBD plugin can do we saw examples about how to create notebooks that analyze and access data that's hosted on the web we talked a little bit about projects and how to organize your work inside notable how you can upload data to those projects and then prompt change gbt to access and analyze fi of the data that's contained in those files in those projects we also learned about how to connect external data connection and then prompt chat TBT to do SQL queries against data and external data connection we also learn a little bit about some of the in-app features of notable like our powerful no code data visualization capability data Explorer or Dex as well as some of the features of a notable Pro Plan that allow you to take your analysis even further with more power and lastly I just want to talk about were some additional resources for help maybe if you're just getting started so please do check out our docs page at docs.notalwood IO that have a lot of information about the notable features as well as a notable chat CPT plugin you can see the list of recent updates as we're making more updates to the plugin right there as well as go to our community page and join the discussion with all the different community members and any questions that you have any future requests that you have please go in there and engage with us and we're more than happy to help so I hope this was helpful and best of luck using the notable chat GPT plugin

Channel: Noteable

Views: 5,773

Keywords:

Id: jrTNvHCdVO8

Length: 15min 14sec (914 seconds)

Published: Wed Aug 23 2023