Great video. You were fine with cleaning as it’s different than formatting. Cleaning=duplicates, typos, unnecessary characters, character vs value (1 vs “1”), field integrity. Formatting= uniformity, readability. Cleaning is how we solve manual data entry errors, formatting is how we ensure the input is readable and avoid output errors.
The term "clean" can be problematic because it implies removing or correcting outlier observations. Since you cannot be certain of what changes an AI might make in the background, it becomes unclear whether the charts reflect the original dataset or a modified, "cleaned" version. This lack of transparency poses a risk, as the modifications might go unnoticed, leading to potential misinterpretations of the data.
That's exactly what I thought. What if even one value is wrong? Depending on the situation this one mistake could be fatal although if you do this for fun I guess it's not an issue.
Very good overview on what's changed with data analysis. So glad you pointed out the cleaning and formatting part because I am sure a lot of people are NOT doing that! I like what they have done with pie charts and graphs regarding changing colors. Thank you for your video! :)
With the scatteplot, you should take into account that your variables are limited and rounded i.e. integers (rounded to the nearest 10 in case of activity). I think you said you had more than 300 data points but 250 of them could have been a sleep rating of 9 and an activity of only 30 and the graph would, misleadingly, look the same. You could roll-over every point but would be better served to introduce some jitter into the points when graphing
The word clean is fine it's letting ChatGPT make the decisions about how to clean it that's problematic. What is it doing with outliers? How about nas? Is it normalizing? How? I would just add a step where you ask it to suggest strategies for cleaning the data based on the dataset and then choose specific things for it to do so that you know what you're actually doing to the data.
Cleaning data can include changing or removing it, especially without more specific boundaries/instructions, and if the desired outcome is to normalize the data, formatting is the correct term. This is a non-trivial, technical difference, not just how you feel about the 2 terms.
Instead of trying to see if there’s a trend in the scattered plot trend, couldn’t you just ask chatgpt if there’s a correlation? Or ask it to draw a regression line?
9:25 if you have asked it to make it more beautiful or more "visulized" it wouldve made it better , im using data analysis 1 year now for visualizing a dataset its better than humans but you need to go in and out a little bit to get the good visualization that top humans can do , for example you couldve used these prompts "make it more colorful" or "make it 3d" or "makethe visualization more advanced" . i rate the update 1/10 since no new ground breaking feature were added and we already couldve used it to create more advanced vizualization .
Hey thanks for the tips! You should join the Ai Foundations Community. You'd be able to help people out a ton with this info my friend. Check it out: swiy.co/aifcommunity
Don't be nervous - what will matter in the future is: What you're most Passionate about AND what Value you bring to your Community or area of expertise. Be innovative, embrace these tools, go places others aren't. It's an opportunity right in front of us. Actually so many opportunities, just choose one and go! I wish you luck.
I was the one to point out the issue with the word "clean", and I appreciate some of the nuanced term tweaks in this video compared to the last, or at least the acknowledgement. It may seem pedantic to many, but part of the problem with all this AI stuff is people not understanding the actual process or meaning of anything produced by the model, and that is an ignorant, dangerous destination.
I hear you but I figure 1) everyone has something to learn and something to teach, 2) young folk doing anything just make me feel old - if I want them to forgive my age I need to forgive their youth, 3) don't be a hater lol ✌️ 🙏 🙇♂️
When you say "interact with your data" does that mean the work that Chat GPT4.0 does is PRIVATE-LOCAL and not on Open AI server? I mean if you are working on something that you want to keep private (like the cure for cancer )) you get my point.
@@ai-foundations I want to sign up for your community. I love learning A.I and one day make it into a career. Does the course inside provide certifications or credit that I can put on my linkend? Thank you Sir.
Sad story is that at the moment it gives you the plot as a png. Even if you ask it to export the graph into an excel file and hand it over to you, it will just insert a picture into the downloadable excel-file. To say it "changed data analysis forever" is quite a huge exaggeration to be honest
@@ai-foundations I think you got me wrong or i didnt explain my problem well enough. I want the plot it gives me as an editable excel-plot. ChatGPT has 0 problems exporting a spreadsheet directly to excel, so there is no need to export is as csv first. Just when you want it to create the plot within the excel-file it is just slapping the png inside that you otherwise have gotten from the chat when directly asking to prompt it.
So a light weight version of #Google #Sheets and graphs is changing the whole Data #Analytics industry…? It looks like the only difference demonstrated is the data source within #ChatGPT… so please, guide me to see the light 💡 here.
I think the biggest thing where this stand out is you can simply prompt and ask questions about your data. Can you ask google sheets a question using natural language and get a response and suggestions for what your data means? LLMs + Data and Analytics is a huge step especially for those who don't know how to interpret data with a confusing tool like sheets and excel. You can ask simple elementary questions and get beautiful insights, trend identification, graphs, etc. Super simple to use and super easy to get started and super powerful having the ability to chat with an LLM about anything relating to your data.
@@ai-foundations from the Side Panel in Sheets within #Workspace, so directly on your data within it, you can use ask Gemini to analyze the data. From #Gemini Advanced you can do this by first uploading your (other) data. But as GPT changed it all, I would think you already knew this. Just like Microsoft, Google is ways ahead when it comes to business usage by integrating GenAI assistants into their offerings, where OpenAI has to build these things / tools around their LLM offering. Also look for: Gemini for Data Scientist and Analysts. It showcases AI and data visualization from a real business (or security) analytics perspective, not just a couple of rows in a spreadsheet.
@@ai-foundations reply is gone. Short: yes, from the side panel in Sheets. And yes, by uploading your data. Have a look at #Gemini for Data Scientists and Analysts for a broader analytics perspective.
Using a decent size report, then this video is no good to you if your just using GPT free version. You reached your limit before it even cleaned the data.
Good work, you guys alwysys make the skool groups so expensive to stsrt loook into the low ticket backend strategy you will make far far more money starting cheaper líke 7 bucks so it's an impulse buy and packing it with people and then upselling more expensive upgrades in the back end, its been proven to make more every time
👉🏼The BEST Ai Community: swiy.co/aifcommunity
👉🏼ChatGPT Mastery Course: hi.switchy.io/chatgpt-mastery-course
This is what Tableau has been designed to do for years before ChatGPT existed.
😊
Great video. You were fine with cleaning as it’s different than formatting. Cleaning=duplicates, typos, unnecessary characters, character vs value (1 vs “1”), field integrity. Formatting= uniformity, readability. Cleaning is how we solve manual data entry errors, formatting is how we ensure the input is readable and avoid output errors.
Good explanation
The term "clean" can be problematic because it implies removing or correcting outlier observations. Since you cannot be certain of what changes an AI might make in the background, it becomes unclear whether the charts reflect the original dataset or a modified, "cleaned" version. This lack of transparency poses a risk, as the modifications might go unnoticed, leading to potential misinterpretations of the data.
yuo wrong🤦♂️🤬👎
True!
all that can be corrected or avoided simply with more detailed instructions, this is not a problem at all.
@@LuisRiveraRomero wrogn🤬👎
That's exactly what I thought. What if even one value is wrong? Depending on the situation this one mistake could be fatal although if you do this for fun I guess it's not an issue.
Very good overview on what's changed with data analysis. So glad you pointed out the cleaning and formatting part because I am sure a lot of people are NOT doing that! I like what they have done with pie charts and graphs regarding changing colors. Thank you for your video! :)
With the scatteplot, you should take into account that your variables are limited and rounded i.e. integers (rounded to the nearest 10 in case of activity). I think you said you had more than 300 data points but 250 of them could have been a sleep rating of 9 and an activity of only 30 and the graph would, misleadingly, look the same. You could roll-over every point but would be better served to introduce some jitter into the points when graphing
The word clean is fine it's letting ChatGPT make the decisions about how to clean it that's problematic. What is it doing with outliers? How about nas? Is it normalizing? How? I would just add a step where you ask it to suggest strategies for cleaning the data based on the dataset and then choose specific things for it to do so that you know what you're actually doing to the data.
Cleaning data can include changing or removing it, especially without more specific boundaries/instructions, and if the desired outcome is to normalize the data, formatting is the correct term. This is a non-trivial, technical difference, not just how you feel about the 2 terms.
@@stephenr85 all the more reason to be specific when using an lm for data analysis!
Agreed. Even ChatGPT admits it sucks at math
Out of the different things that can be called data analysis, this is at most descriptive and visualizing.
Instead of trying to see if there’s a trend in the scattered plot trend, couldn’t you just ask chatgpt if there’s a correlation? Or ask it to draw a regression line?
Yes you could! Great thinking man
Yep. Relationship would be a better term. Trend implies something changing over time
@@yoagcur yes you’re right, my mistake. English is not my first language, so my vocabulary is probably not as precise as it should be 😅
Is there any way to use ChatGPT to analyze verbatim and qualitative datas ?
Keep slaying bro, 🔥🔥 I appreciate all your teachings.
Excluding user input. Is the data live, or is it pulling from old sources?
9:25 if you have asked it to make it more beautiful or more "visulized" it wouldve made it better , im using data analysis 1 year now for visualizing a dataset its better than humans but you need to go in and out a little bit to get the good visualization that top humans can do , for example you couldve used these prompts "make it more colorful" or "make it 3d" or "makethe visualization more advanced" . i rate the update 1/10 since no new ground breaking feature were added and we already couldve used it to create more advanced vizualization .
Hey thanks for the tips!
You should join the Ai Foundations Community.
You'd be able to help people out a ton with this info my friend.
Check it out:
swiy.co/aifcommunity
It seems the updated (cleaned and formatted) spreadsheet is not downloaded. Would there be a reason the updated data sheet does not download?
@ai-foundation how are you able to generate the spreadsheets in chatgapt-4o? I have susbcribed but my sample data comes in text format.
I use GPT Chat, the paid version.. How can you make a visual representation of the data interactive and not just an image?
thanks for the video. did you notice your browser slowing down to a crawl after a few messages with code interpreter? how did you solve that?
I haven't noticed any slowing down of my browser... that's odd
Ok, but it can make dumb mistakes leading you to wrong conclusions, so you can never be sure of the acuracy of the analysis..
Not everytime and even humans mistakes are dumb mistakes
Wow…people voluntarily wear a ring that gives a tech company insight into everything in your life. I’d be chucking it into Mt Doom.
If someone is learning Data Analytics currently, should they be worried about this?
I think they should be excited about this. They will have all the pro knowledge of Data Analytics and an assistant that makes their life 10x easier.
Don't be nervous - what will matter in the future is:
What you're most Passionate about AND what Value you bring to your Community or area of expertise.
Be innovative, embrace these tools, go places others aren't. It's an opportunity right in front of us.
Actually so many opportunities, just choose one and go!
I wish you luck.
Yes, now someone who was in the business 10-20 years and now is redundant, will get retrained and won't give you the position you want
@@Xarabek Actually I have around 3.5 years of exp, what should I do to be safe.
@@wolf8605 depends what you like really, going into ML is a good oath considering you already have experience
This is fire content.
Thank you so much.
Should I be using the ChatGPT's Data Analyst GPT instead of the ChatGPT interface? What is the difference??
I was the one to point out the issue with the word "clean", and I appreciate some of the nuanced term tweaks in this video compared to the last, or at least the acknowledgement. It may seem pedantic to many, but part of the problem with all this AI stuff is people not understanding the actual process or meaning of anything produced by the model, and that is an ignorant, dangerous destination.
Yes, thanks for pointing it out!
Wait where the data taken from itself, I mean the source? And if it is generated like not-real data, how can it be bring into report?
I hear you but I figure 1) everyone has something to learn and something to teach, 2) young folk doing anything just make me feel old - if I want them to forgive my age I need to forgive their youth, 3) don't be a hater lol
✌️ 🙏 🙇♂️
Amazing, thanks for the learnings
Very useful keep going
Is it valid for only premium members
Great content!
Can I ask you the Top 50 companies dataset? Thnks
Let’s gooo!
Yessir!
@@ai-foundations This video is popping brother. So much value and love from you in this one. ♥
Is data entry a good way to earn extra cash ?
When you say "interact with your data" does that mean the work that Chat GPT4.0 does is PRIVATE-LOCAL and not on Open AI server? I mean if you are working on something that you want to keep private (like the cure for cancer )) you get my point.
Claude 3.5: "Hold my beer"
How much to contract you? Seriously?
Great job buddy !
I like it , very good presentation
Thanks a lot. Great video. May God bless you
Thank you sir! You are awesome!❤
Thank you for the kind words and feedback Aaron!
Why doesn't ChatGPT read online videos anymore?
Not sure, run python on make.com and pull TH-cam transcript api
Likely due to the fact that the video doesn’t have a transcript or subtitles. Happens with Chat GPT Custom GPTs and Gemini.
Great video!
Well done!
Amaging video ❤
Great thanks 🙏
Of course!
@@ai-foundations I want to sign up for your community. I love learning A.I and one day make it into a career. Does the course inside provide certifications or credit that I can put on my linkend? Thank you Sir.
We expose ourselves into gpt hallucinations while cleaning. I suggest use that with caution. 😐
Is data analytics jobs cooked??
167k vs
There are lies, damn lies and statistics
Sad story is that at the moment it gives you the plot as a png. Even if you ask it to export the graph into an excel file and hand it over to you, it will just insert a picture into the downloadable excel-file.
To say it "changed data analysis forever" is quite a huge exaggeration to be honest
Export as a CSV and import it to excel.
@@ai-foundations I think you got me wrong or i didnt explain my problem well enough. I want the plot it gives me as an editable excel-plot. ChatGPT has 0 problems exporting a spreadsheet directly to excel, so there is no need to export is as csv first. Just when you want it to create the plot within the excel-file it is just slapping the png inside that you otherwise have gotten from the chat when directly asking to prompt it.
This seems it can analyze and visualize basics of analysis only
Thank you
Great video like this
Can one become data analyst using chatgpt only?? Plz reply
Holy crap everything changed again?! That's the 10th time this week!
This industry changes fast. It really does.
@@ProductiveDude but these click bait titles never do
💙 Thanks 💙
Good studf here ... ty
So a light weight version of #Google #Sheets and graphs is changing the whole Data #Analytics industry…? It looks like the only difference demonstrated is the data source within #ChatGPT… so please, guide me to see the light 💡 here.
I think the biggest thing where this stand out is you can simply prompt and ask questions about your data. Can you ask google sheets a question using natural language and get a response and suggestions for what your data means? LLMs + Data and Analytics is a huge step especially for those who don't know how to interpret data with a confusing tool like sheets and excel. You can ask simple elementary questions and get beautiful insights, trend identification, graphs, etc. Super simple to use and super easy to get started and super powerful having the ability to chat with an LLM about anything relating to your data.
@@ai-foundations from the Side Panel in Sheets within #Workspace, so directly on your data within it, you can use ask Gemini to analyze the data. From #Gemini Advanced you can do this by first uploading your (other) data. But as GPT changed it all, I would think you already knew this. Just like Microsoft, Google is ways ahead when it comes to business usage by integrating GenAI assistants into their offerings, where OpenAI has to build these things / tools around their LLM offering. Also look for: Gemini for Data Scientist and Analysts. It showcases AI and data visualization from a real business (or security) analytics perspective, not just a couple of rows in a spreadsheet.
@@ai-foundations reply is gone. Short: yes, from the side panel in Sheets. And yes, by uploading your data. Have a look at #Gemini for Data Scientists and Analysts for a broader analytics perspective.
@@ai-foundationsI can ask power bi for example, this functionality has been available a long time ago.
Using a decent size report, then this video is no good to you if your just using GPT free version. You reached your limit before it even cleaned the data.
This kid says chatgpt and average of 23.5 times per video. This statistics was given to me by chatgpt.
A useful video with nice explanations. Keep up the good work. I also create videos on AI-related technologies. Thank you.
Nice!
Thank you :)
I see a lot of data visualization in the video but no real data analysis.
I hate learning stuff from people who look like they’ve just finished playing a middle school basketball game.
This mentality >>> 🔥 Great wisdom Todd!
@@ai-foundationsI disagree
I found this video is very value, before you say a bad thing show-us what better content you offer!.
@@anonymouswikileaks2451 i agree with you
That’s so rude. What’s wrong with you? Be thankful people will take the time out of their day to post great content for free so you can learn.
Too basic to be useful for the moment. Perhaps convenient for a quick visual on a very simple dataset.
Fair enough!
Speaking very more then showing
‘This could replace Google sheets in the future.’ Bro plz. This makes me stop the vid. Honest feedback
This kid is making a TH-cam channel like everyday- it’s pathetic- crime, court, reality he is making a fool of us
dude SHUT UP i cant heaer myself think
You're a nerd 😂
🤓
Good work, you guys alwysys make the skool groups so expensive to stsrt loook into the low ticket backend strategy you will make far far more money starting cheaper líke 7 bucks so it's an impulse buy and packing it with people and then upselling more expensive upgrades in the back end, its been proven to make more every time
Full guide ... inside the paid group