I'm a teacher learning data analysis/data scientist and one thing I found working with ChatGPT that reminds me of teaching is that with a class of students, you can't just give instructions to a class activity and then sit back and wait for them to finish. You have to make sure everyone understood but observing their work/progress. And that's a lot like ChatGPT, you can't just ask it to do something and then blindly assume what it gives you is correct. You still need to check what it produces to make sure the two of you are on the same page.
Do/will chatbots create summaries that reflect the interrogators bias? As an example, the differences in the job stat differences were not that significant, but the limitations were explained. if a bot reflects the interrogator"s opinion then it could overweight the case. A lot of the problems I see is that people don't understand their data, (especially the limitations of it, the cost/benefits of getting better data), and don't understand numbers enough to ask the right questions and sense they have the wrong answers
The iterative process needed to get GPT3/4 to output quality work is an excellent teaching tool. In reality it’s similar to working with a development team where you have to sort of organize and coordinate multiple strings of thought so that you can create a final product. And the better you are at describing what you need, revising what you need and conveying those needs to the developers, the better product you’ll get at the end. I don’t think everyone can just jump into GPT4 and get great results, not right away, I think you need the right mindset to work iteratively which isn’t even pervasive among actual developers. At the end of the day GPT4 won’t make someone a great developer but great developers who use GPT4 will likely surpass developers who shun away from it.
My take away is that, when the analysis that you want to perform is complex, it takes way more time to come up with the right prompt and test the reliability of the results, than to write simple python code that visualizes everything reliably. Sometimes its easier to code than doing trial and error with prompts :) . But chatGPT seems to be doing great at generating quick results from a lot of unstructured sources when you need quick insights.
I think you are the first to explain in detail what these plug-in stuffs are about with Chatgpt. I've been using Chatgpt for a time now and it's been so great. Thanks for more clarity
As a data scientist/ semiconductor researcher, I am very excited about this sort of “automation”. GPT4 is like a calculator on steroids, I can get my work done faster and better. And no it’s not always right but that’s where my knowledge as an SME comes into play and eventually AI will get better so it doesn’t matter too much. Secondly this stuff lets me spend way more time focusing on creative thought problems and less on brute forced analysis. And that freedom of time is invaluable because that’s where the leaps in tech come from, people thinking up cool stuff and then making it. AI helps give me more time than I’ve ever had before and I can use that time to advance my field.
Plugins look pretty awesome to try! Privacy however has been of utmost concern with financial and healthcare companies even if ChatGPT and plugins produce the best analytics results. You simply cannot do any internal data analysis until legal and engineering have signed off or until you build an in-house pipeline that does not communicate externally, and all of that requires resources. Code generation seems like the best use case as that is largely input agnostic and does not reveal any confidential data.
Let's be honest, most of the data we work with are not that confidential. I really need to get fluent in the usage of these plugins. Any resources you can recommend?
It makes sense that the data analyst would like using chat GPT as a tool, but people in the field would be apprehensive about it's effects on the field. Quite simply it's a tool that increases the productivity of individuals, so the individuals using the tool will find their job easier, but at the same time it increases competition within the field because as each person becomes more productive the field becomes more competitive and there will be fewer jobs or maybe even more demands on the individuals. If you were the only person who used chat GPT and your managers did not know about it, yeah chat GPT would be a great tool. Just like if you were in a bar fight and you were the only person that had a gun, you would be at a distinct advantage. But if you're in a bar fight and suddenly everybody has guns well then it turns from a bar fight into a massacre
Very thorough video as always! Also the thing about why it gives different values compared to what you see in the pdf itself, from what i read about how generative ai and tools like chatgpt work is they try to *guess* what the true response *should* be, that’s why in this context they’re not very good at working with numbers. For example if you give it a simple multiplication of two numbers it gives something that *looks* right even though it isn’t most of the time.
Here is my take as an accountant. Ai tools will hit the tech department first and then slowly bleed into other departments with add ons to their current software tools. For example, no accountant I know will use this but include it as a free addon to their current erp and it will be used. It will take time but as these tools will eventually find their way into everything. Will it take your job, probably no. Will it change the way you do your job probably yes.
As long as AI tools are not implemented in companies, these will not require employees to have that skill. It all boils down to privacy issues, as you mentioned. That will take some time. If I were to guess, that's a frame of 1-2 years, depending on the law, which is being established right now (in Europe at least). edit: that applies to big companies, smaller ones might get into that earlier.
it won't ever exist. It's like having the requirement of "knowing how to google" which is nothing more than just a joke if you're putting it into the job description. developers might already be using it in their job with or without the companies knowing.
*_"If it ain't broke don't broke, don't fix it"_* That's the mentality of the companies I worked for so it's not surprising that the survey showed few companies requiring AI related knowledge.
I thougt about this just yesterday!! And today you post a video. You're like a guru to me. I love your videos and try to insert this in my life as a student.
Exceptional analysis Luke. 👏🏻👏🏻 I am from a non tech background and worked in the customer service industry for 4 years and finally entering the world of data science. I am doing a course where they help you to upskill yourself even if you're not from a tech background. I am very positive about AI and data science going hand in hand. I did a real world analysis of the students in my batch (we have a mixed batch, consists of people already working in the data field and people who are freshers in data), I found that the freshers are more inclined towards using AI tools to perform day to day tasks and even learning new concepts. The people who are already into data, have heard about AI tools but the new people entering into data are actually using AI tools to get their work done. Even for me, I use chatgpt to debug any python code for me and to find alternate ways to solve a problem as it reduces dependency on others for easy problems. I have noticed a bias in the freshers, that is, they believe if they aren't good with AI tools, they will have a hard time finding a job in data industry. Would love to know your thoughts on this. 😃
It’s actually my most comfortable shirt. I need to get more from you “LukeGPT” is one plan above the Plus version. I’ll see if I can get you on the invite list 😜
I totally relate to the moment when you got some graphs out and think "Hm.. I don't like this!" 🤣Love this video and the analyses! 💖 The Noteable plugin seems pretty solid! We're definitely facing a big time as data professionals RN
It's too soon for this analysis ... companies are 99% managed by people who don't know what they don't know about AI potential - of course the geeky roles are more excited than the people who decide wether AI is desirable / impactful for the corporate world, FOR NOW. If you watched the video and reading this - don't interpret Luke's video as a "don't care too much about AI... maybe even disregard it" - I would probably recommend the opposite regardless of every (kinda expected giving the timing of it) insight that resulted from this. In other words - I'm not a programmer, but I been in market intelligence and strategic planning in a top #1 company of an X big LATAM country; and I'll be using this channel A LOT to learn about becoming an average data scientist with AI kickers everywhere I can. Thanks for existing Luke!!!
Thanks for your video Luke. I use chat GPT a lot, but only until certain level. Because the AI responds you with extreme confidence and sometimes the responses are wrong, because it skips part of the logic or the data. It is great for a preliminar view or to accelerate insights, but in a serious job it has to be double checked. Another thing, that you even mention in your video, is the high concern about data privacy. If I may ask you something. There are a lot of companies asking for deep knowledge in ETL processes and Data Warehousing. I have seen many offers asking for snowflake, AWS, BigQuery, Spark, knowledges of Data Lakes, etc. I completed many of the courses that you've recommended for the basic tools (SQL, Tableau, Excel, the Data Analytics from Google) and they were all very good. That's why I'd like to aks you if you have any recommendation for me to get to the next level of the ETL process. Thank and all the best.
I'd really recommend focusing know on projects to build up that experience with ETL. That will produce tangible results that you can then showcase to employers and give real world examples of how you've built up your ETL skills. For my projects, i'm learning a lot along the way and it's relevant to my project so it helps with maintaining motivation
I really like how you choose your application as data analysts, thanks for data set although it's not conclusive somehow it give me a picture of what's its gonna be a year from now. I'm really worried about how Ai will affect this job in the future and it gives me somehow a better perspective.
The Notable API plug in to Chat GPT 4 + gives more .Furthermore, Chat GPT clearly states that it is experimental and is a work in progress, constantly learning. Nit picking right now is foolhardy . It will only get exponentially better . Hype now but it’s here and won’t go away. Privacy around data sources might be a stumbling block but you just know businesses will overcome these because of cost and speed benefits.
Awesome video, Luke! With this whole obscurity around the future of jobs, do you think studying towards becoming a data engineer will be better than being a data analyst
I don't think there is much obscurity with jobs RN, this is just another tool to make us more efficient. Whether data engineer or anlayst it's really up to your own interests
AI creators are big tech companies like Microsoft and Google. But AI users will be small and medium tech industries. So many companies will avoid using AI tools for data privacy issues. As people who have all the data of the world will be more profitable then the one who is going to use it and by unknowingly sharing the critical information with the big tech company, could be a costly business decision or hamper the growth of a company.
🎯 Key Takeaways for quick navigation: 00:00 👩💼 Chat GPT has been helping the speaker save time in data analysis by providing quick data insights and visualizations without coding. 02:04 📚 Chat GPT, running on OpenAI's GPT-4 model, is a powerful AI chatbot for data analysis, but advanced features require the pro version with access to plugins. 05:01 🔍 Chat GPT's limitations include being trained only up to September 2021, but plugins like the Bing plugin help access recent information for analysis. 07:33 👍 Around 77% of data professionals have a favorable opinion towards using AI tools in their job workflows. 10:18 💼 Less than 2% of job postings for data science roles request AI-related skills in their job descriptions, raising questions about the gap between individual interest and company implementation. Made with HARPA AIha
Hi Luke 👋 great content! I am a data scientist who wants to start a youtube channel too. I am just wondering that are you editing and organizing your videos or somebody else is doing it for you?
I'm currently editing all my long-form videos (I had an editor for a bit for my short-form). If you like editing then I recommend doing it yourself, but if you're not into it then I recommend just hiring an editor
I work for a relatively big company. 6000 employees 4 biljoen revenue. My company is very conservative adapting new technologies. I’m sure the company I work for isn’t unique in that sense. Will they adapt in the future….probably they will. But that will only move the field for data analytics further in the organization. There are so many use cases that it will take a very long time implement all of those. So should you be worried…nah….should you ignore it?…also definitely not.
I’m using the Noteable plugin for this as well. You just provide the link and tell it to connect. If it’s giving you issues just tell it specifically to use the ‘gspread’ python library.
Hi Luke, as per you analysis and with the rise in AI tools like these and plugins such as the code interpreter, do you still think it is worth becoming a data analyst? 🤓🤔
Very interesting, I think that the economy situation we are living in has to do with the companies response towards AI. Maybe this will dramatically chague in the near future.
Hello Luke, what is your opinion on using AI in Data science jobs from business perspective and most analysis has to be done based on historical data of customers or any companies data then there is a chance that data we are feeding to AI can be risky. While using chatgpt, what are do's and don'ts especially people working in companies.
It could be fabricating the data because, 1: it couldn't read it properly. 2: it ran out of tokens and couldn't process that data. Those are my two theories.
Such an awesome video, I am just getting into data analytics as a career and am worried if Im wasting my time as these AI tools may well do the task 10x better in a millisecond of the time but it seems that's not about to happen soon. In your video between seconds 7 and 13 you linked a spreadsheet to chatpgpt. It went by in a flash, how do you actually step by step link them?
4 out of 5 is an A+? Confirmed Luke plays JRPGs and knows S tier is the highest. One of the frustrating things I hear about are those companies that have blocked these AI tools from being used because they deal with sensitive data and employees can’t be trusted to use it properly.
Haha you know it’s s-tier! And yeah companies have to start thinking of what there workaround is. Just saw the other day Google is banning use of Bard for Google employees to input their data into. So craxy
What do you think about the Code Interpreter in ChatGPT (Able to generate charts, maps, data viz, clean datasets)? There're tweets and news explaining that "anyone can be a data analyst now" or "anyone can be hired a data analyst for only $20/month". I've been studying for 3 months and increasingly afraid that in the future data analysts are not needed, am I too late to become a data analyst?
Can you tell us how to use these plugins to make these analyses and visualizations directly in ChatGPT through plain text prompts? This would be the push I needed to move from the free to paid ChatGPT Plus account.
Great question. It really depends on the prompting IMO. If I keep it on track with descriptive prompts I feel it performs much better and don’t really have to worry about hallucinations.
For me, no. The data I have I don't have security concerns with. And I think many fall into this category; But others that may have confidential data (HIPPA data, etc.); yeah they def shouldn't put that in here
@@LukeBarousse great, thanks! Do you think anonymizing HIPPA data is sufficient? When I said share confidential data, I meant more like a company's internal proprietary data. Everyone I've asked will side with caution and say "Be careful what you share with ChatGPT, if you wouldn't want the data published on the internet then don't share it" ...but then I don't see how we can unlock it's full potential. Do you have any additional thoughts?
I was wondering if someone could help me plan out a “Roadmap to first Data Analytics Job”. I’ve seen so many different videos, each of them recommending different paths. Does someone have a preferred “path” to take for someone starting with little to no data experience?
🎯 Key Takeaways for quick navigation: 00:00 📊 *Using ChatGPT for Data Analysis* - Demonstrates using ChatGPT for data analysis tasks, significantly reducing time compared to manual methods. - Utilizes ChatGPT to connect to a spreadsheet, analyze job postings, and generate visualizations like bar graphs and salary calculations. 01:23 🤖 *Understanding AI's Impact on Data Jobs* - Explores the impact of AI on both individuals and businesses within the realm of data jobs. - Outlines a two-step approach involving data research and analysis using ChatGPT to investigate sentiments and effects of AI on data science roles. 02:31 💼 *Exploring ChatGPT Capabilities* - Provides insights into ChatGPT, an AI model by OpenAI, highlighting its effectiveness in data analysis. - Discusses the benefits of the pro version, including access to plugins for enhanced functionality. 04:05 🔍 *Searching for Information on AI Impact* - Discusses limitations of ChatGPT in accessing real-time data and the use of plugins like Bing for internet searches. - Utilizes ChatGPT to analyze PDF documents and surveys, extracting insights about AI's impact on data-related tasks. 05:55 📊 *Analyzing Developer Sentiment Towards AI* - Examines survey results from developers regarding their sentiment towards AI in the workplace. - Utilizes ChatGPT to summarize and analyze survey findings, revealing positive sentiments among data professionals towards AI technology. 08:01 💼 *Assessing Business Perspectives on AI* - Explores how businesses perceive AI's impact on economic output and workforce dynamics. - Discusses potential challenges such as privacy concerns while emphasizing the benefits of AI adoption for companies. 09:10 📊 *Analyzing Job Postings for AI Skills* - Utilizes ChatGPT and plugins for data analysis, focusing on job postings to understand the demand for AI skills. - Faces challenges such as dataset size limitations and sampling issues but generates insights into the prevalence of AI-related skills in job descriptions. 11:45 🔍 *Exploring AI Tools in Job Descriptions* - Investigates the inclusion of AI-related tools in job descriptions, revealing low percentages of job postings requiring AI skills. - Analyzes salary differentials among job titles, highlighting discrepancies in compensation for roles requiring AI expertise. 13:09 💡 *Concluding Remarks on AI Implementation* - Summarizes findings indicating a disparity between individual interest in AI and its adoption in job postings. - Calls for further exploration into the reasons behind this disconnect and suggests future research directions. Made with HARPA AI
Conclusion on hiring did not age well. ML and Gen AI skills hiring is defining the tech economy in 2024 and probably through 2027, at least, though of course there will be biggish layoffs, as well. Incredible workflows though, thank you!
the hallucinations are such a bummer. hard to have confidence when it routinely makes stuff up. qc nightmare. i know it will get better, but for business use cases it's still pretty sketch at the moment. very much swim at your own risk.
Mask the sensitive data. If I use it for work, I change names of things with the name of the client to the word "client" for example, if there is one client. If there are personal ID numbers, just split them and concatenate them with a text string of something unique for that row of data, that is not sensitive data, so you can revert it easily.
![[TH] 2024 PMGC League | Survival Stage วันที่ 1 | PUBG MOBILE Global Championship](http://i.ytimg.com/vi/FVFT-PFwU80/mqdefault.jpg)
I'm a teacher learning data analysis/data scientist and one thing I found working with ChatGPT that reminds me of teaching is that with a class of students, you can't just give instructions to a class activity and then sit back and wait for them to finish. You have to make sure everyone understood but observing their work/progress. And that's a lot like ChatGPT, you can't just ask it to do something and then blindly assume what it gives you is correct. You still need to check what it produces to make sure the two of you are on the same page.
Totally agree. I also like to break it down into steps and have it do it one by one so it’s going through the correct process
You can also ask it to provide steps why it came to some conclusion and weirdly that should reduce hallucinations. This field is in such a flux now 😁
From one teacher and CGPT user to another, well spoken! 🤣🙌🏾
Do/will chatbots create summaries that reflect the interrogators bias?
As an example, the differences in the job stat differences were not that significant, but the limitations were explained.
if a bot reflects the interrogator"s opinion then it could overweight the case.
A lot of the problems I see is that people don't understand their data, (especially the limitations of it, the cost/benefits of getting better data), and don't understand numbers enough to ask the right questions and sense they have the wrong answers
The iterative process needed to get GPT3/4 to output quality work is an excellent teaching tool. In reality it’s similar to working with a development team where you have to sort of organize and coordinate multiple strings of thought so that you can create a final product. And the better you are at describing what you need, revising what you need and conveying those needs to the developers, the better product you’ll get at the end. I don’t think everyone can just jump into GPT4 and get great results, not right away, I think you need the right mindset to work iteratively which isn’t even pervasive among actual developers. At the end of the day GPT4 won’t make someone a great developer but great developers who use GPT4 will likely surpass developers who shun away from it.
My take away is that, when the analysis that you want to perform is complex, it takes way more time to come up with the right prompt and test the reliability of the results, than to write simple python code that visualizes everything reliably. Sometimes its easier to code than doing trial and error with prompts :) . But chatGPT seems to be doing great at generating quick results from a lot of unstructured sources when you need quick insights.
I am by any metric new to data science and ChatGPT has been invaluable, particularly in my learning of SQL, Python, and Google Cloud and coding.
Same here. It's been very helpful
I think you are the first to explain in detail what these plug-in stuffs are about with Chatgpt.
I've been using Chatgpt for a time now and it's been so great.
Thanks for more clarity
As a data scientist/ semiconductor researcher, I am very excited about this sort of “automation”. GPT4 is like a calculator on steroids, I can get my work done faster and better. And no it’s not always right but that’s where my knowledge as an SME comes into play and eventually AI will get better so it doesn’t matter too much. Secondly this stuff lets me spend way more time focusing on creative thought problems and less on brute forced analysis. And that freedom of time is invaluable because that’s where the leaps in tech come from, people thinking up cool stuff and then making it. AI helps give me more time than I’ve ever had before and I can use that time to advance my field.
This is where I'm at too. Different domain, but same theme
Plugins look pretty awesome to try!
Privacy however has been of utmost concern with financial and healthcare companies even if ChatGPT and plugins produce the best analytics results. You simply cannot do any internal data analysis until legal and engineering have signed off or until you build an in-house pipeline that does not communicate externally, and all of that requires resources.
Code generation seems like the best use case as that is largely input agnostic and does not reveal any confidential data.
Let's be honest, most of the data we work with are not that confidential. I really need to get fluent in the usage of these plugins. Any resources you can recommend?
It makes sense that the data analyst would like using chat GPT as a tool, but people in the field would be apprehensive about it's effects on the field. Quite simply it's a tool that increases the productivity of individuals, so the individuals using the tool will find their job easier, but at the same time it increases competition within the field because as each person becomes more productive the field becomes more competitive and there will be fewer jobs or maybe even more demands on the individuals. If you were the only person who used chat GPT and your managers did not know about it, yeah chat GPT would be a great tool. Just like if you were in a bar fight and you were the only person that had a gun, you would be at a distinct advantage. But if you're in a bar fight and suddenly everybody has guns well then it turns from a bar fight into a massacre
It won't increase competition but will decrease demand like photography
Very thorough video as always! Also the thing about why it gives different values compared to what you see in the pdf itself, from what i read about how generative ai and tools like chatgpt work is they try to *guess* what the true response *should* be, that’s why in this context they’re not very good at working with numbers. For example if you give it a simple multiplication of two numbers it gives something that *looks* right even though it isn’t most of the time.
Here is my take as an accountant. Ai tools will hit the tech department first and then slowly bleed into other departments with add ons to their current software tools. For example, no accountant I know will use this but include it as a free addon to their current erp and it will be used. It will take time but as these tools will eventually find their way into everything. Will it take your job, probably no. Will it change the way you do your job probably yes.
As long as AI tools are not implemented in companies, these will not require employees to have that skill. It all boils down to privacy issues, as you mentioned. That will take some time. If I were to guess, that's a frame of 1-2 years, depending on the law, which is being established right now (in Europe at least).
edit: that applies to big companies, smaller ones might get into that earlier.
it won't ever exist. It's like having the requirement of "knowing how to google" which is nothing more than just a joke if you're putting it into the job description. developers might already be using it in their job with or without the companies knowing.
*_"If it ain't broke don't broke, don't fix it"_*
That's the mentality of the companies I worked for so it's not surprising that the survey showed few companies requiring AI related knowledge.
I thougt about this just yesterday!! And today you post a video. You're like a guru to me. I love your videos and try to insert this in my life as a student.
Exceptional analysis Luke. 👏🏻👏🏻
I am from a non tech background and worked in the customer service industry for 4 years and finally entering the world of data science. I am doing a course where they help you to upskill yourself even if you're not from a tech background. I am very positive about AI and data science going hand in hand. I did a real world analysis of the students in my batch (we have a mixed batch, consists of people already working in the data field and people who are freshers in data), I found that the freshers are more inclined towards using AI tools to perform day to day tasks and even learning new concepts. The people who are already into data, have heard about AI tools but the new people entering into data are actually using AI tools to get their work done. Even for me, I use chatgpt to debug any python code for me and to find alternate ways to solve a problem as it reduces dependency on others for easy problems. I have noticed a bias in the freshers, that is, they believe if they aren't good with AI tools, they will have a hard time finding a job in data industry. Would love to know your thoughts on this. 😃
Thank you Luke. A great master class on how to derive insights. You really are an inspiration to all the data analysts "in the making" of the World.
Elite clothing choice at 1:18. Also loved the new green screen Luke in my Chat GPT window. How can I get that feature?
It’s actually my most comfortable shirt. I need to get more from you
“LukeGPT” is one plan above the Plus version. I’ll see if I can get you on the invite list 😜
@@LukeBarousse please do! I have JEE-PT coming out soon
Put me on the waitlist!!
I totally relate to the moment when you got some graphs out and think "Hm.. I don't like this!" 🤣Love this video and the analyses! 💖 The Noteable plugin seems pretty solid! We're definitely facing a big time as data professionals RN
Ha! Glad I’m not alone on the graphs 🙌🏼 😂
one thing that is needed badly but missing is the ability to switch between models within same chat session!
Really great run down and video. Interesting to see different ways you used this tool, thanks Luke!
You have some mad editing skills
There's a reason I went hard at the ML / AI development and NLP route.
And this is exactly why.
I am a data analyst and I found out that before chat gpt I was Klark Kent. Now I am Superman 🤞💪. It is a great boost to my job.
It's too soon for this analysis ... companies are 99% managed by people who don't know what they don't know about AI potential - of course the geeky roles are more excited than the people who decide wether AI is desirable / impactful for the corporate world, FOR NOW. If you watched the video and reading this - don't interpret Luke's video as a "don't care too much about AI... maybe even disregard it" - I would probably recommend the opposite regardless of every (kinda expected giving the timing of it) insight that resulted from this. In other words - I'm not a programmer, but I been in market intelligence and strategic planning in a top #1 company of an X big LATAM country; and I'll be using this channel A LOT to learn about becoming an average data scientist with AI kickers everywhere I can. Thanks for existing Luke!!!
You are teaching me a lot thanku so much sir even I am using chatgpt for data analysis
Thanks for your video Luke.
I use chat GPT a lot, but only until certain level. Because the AI responds you with extreme confidence and sometimes the responses are wrong, because it skips part of the logic or the data. It is great for a preliminar view or to accelerate insights, but in a serious job it has to be double checked. Another thing, that you even mention in your video, is the high concern about data privacy.
If I may ask you something.
There are a lot of companies asking for deep knowledge in ETL processes and Data Warehousing. I have seen many offers asking for snowflake, AWS, BigQuery, Spark, knowledges of Data Lakes, etc.
I completed many of the courses that you've recommended for the basic tools (SQL, Tableau, Excel, the Data Analytics from Google) and they were all very good. That's why I'd like to aks you if you have any recommendation for me to get to the next level of the ETL process.
Thank and all the best.
I'd really recommend focusing know on projects to build up that experience with ETL. That will produce tangible results that you can then showcase to employers and give real world examples of how you've built up your ETL skills. For my projects, i'm learning a lot along the way and it's relevant to my project so it helps with maintaining motivation
I really like how you choose your application as data analysts, thanks for data set although it's not conclusive somehow it give me a picture of what's its gonna be a year from now. I'm really worried about how Ai will affect this job in the future and it gives me somehow a better perspective.
Love this format! Great video my friend 😊
Hi Luke, I like your content very informative and interesting. Keep up the good content.
You're the man Luke! Great video!
Thanks Avery!! Hope to get to see ya again soon my dude 🙌🏼
1:04 1 minute, that's now but next year it could be 1 second
I think thats all good. Just means analysts can be way more efficient using ai, while businesses dont even know theyre using it
Thanks for posting Luke. Hopefully we are not screwed
How do you make a video about Data Analysis?
Random TH-cam Data Analysts: I just talk about it.
Luke: I use my Data Analyst skills to make the video.
🙌🏼🙌🏼🙌🏼
Love the topic & Content!
Keep sharing the good work (pls no ads! take sponsors instead), you can cross 1M by end of this year
Can you make a video about doing a analyst project with AI help and without. I am really curious what the difference will be with your skill set
You make my life easier, thanks brother ❤️
I hit that like button because I enjoy watching this data nerds videos ❤
With this fast in ai development .
all repetitive work using computers can be automated within 3 years
Luke, amazing insights, thanks for the video!
🙌🏼
This is cool as an analyst I think itll be a requirement to know how to use these tools
The Notable API plug in to Chat GPT 4 + gives more .Furthermore, Chat GPT clearly states that it is experimental and is a work in progress, constantly learning. Nit picking right now is foolhardy . It will only get exponentially better . Hype now but it’s here and won’t go away. Privacy around data sources might be a stumbling block but you just know businesses will overcome these because of cost and speed benefits.
Agree it's experimental; my goal at "nitpicking" is to help provide feedback with the development.
Awesome video, Luke!
With this whole obscurity around the future of jobs, do you think studying towards becoming a data engineer will be better than being a data analyst
I don't think there is much obscurity with jobs RN, this is just another tool to make us more efficient. Whether data engineer or anlayst it's really up to your own interests
The question to Bard at 4:38, if you use ‘generative AI’ instead of AI, then it works. After all, data scientists are already working with AI.
Some very interesting analysis and conclusions, appreciate the video, thanks.
AI creators are big tech companies like Microsoft and Google.
But AI users will be small and medium tech industries. So many companies will avoid using AI tools for data privacy issues.
As people who have all the data of the world will be more profitable then the one who is going to use it and by unknowingly sharing the critical information with the big tech company, could be a costly business decision or hamper the growth of a company.
🎯 Key Takeaways for quick navigation:
00:00 👩💼 Chat GPT has been helping the speaker save time in data analysis by providing quick data insights and visualizations without coding.
02:04 📚 Chat GPT, running on OpenAI's GPT-4 model, is a powerful AI chatbot for data analysis, but advanced features require the pro version with access to plugins.
05:01 🔍 Chat GPT's limitations include being trained only up to September 2021, but plugins like the Bing plugin help access recent information for analysis.
07:33 👍 Around 77% of data professionals have a favorable opinion towards using AI tools in their job workflows.
10:18 💼 Less than 2% of job postings for data science roles request AI-related skills in their job descriptions, raising questions about the gap between individual interest and company implementation.
Made with HARPA AIha
Oh I've been tempted
11:21 Like a parent saying no xbox till after you clean your room xD
Great video Luke!
Hi Luke 👋 great content! I am a data scientist who wants to start a youtube channel too. I am just wondering that are you editing and organizing your videos or somebody else is doing it for you?
I'm currently editing all my long-form videos (I had an editor for a bit for my short-form). If you like editing then I recommend doing it yourself, but if you're not into it then I recommend just hiring an editor
@@LukeBarousse thank you for your response 🙏🏻
I work for a relatively big company. 6000 employees 4 biljoen revenue. My company is very conservative adapting new technologies. I’m sure the company I work for isn’t unique in that sense. Will they adapt in the future….probably they will. But that will only move the field for data analytics further in the organization. There are so many use cases that it will take a very long time implement all of those. So should you be worried…nah….should you ignore it?…also definitely not.
Thank you 👍
Hello luke as a non-technical background how we learn data science step by step process???
I’ve been searching and searching… what plug-in allows you to reference data from a google sheet?
I’m using the Noteable plugin for this as well. You just provide the link and tell it to connect. If it’s giving you issues just tell it specifically to use the ‘gspread’ python library.
I love the idea of a green screen special this moment 3:47 made me laugh 😅
Hi Luke, as per you analysis and with the rise in AI tools like these and plugins such as the code interpreter, do you still think it is worth becoming a data analyst? 🤓🤔
Yes, no time like the present!
Can you make a video on how to get international remote job as a data analyst/scientist
What about data security of connecting your database to this thing? Is it safe to connect work or even HIPPA data to this?
Very interesting, I think that the economy situation we are living in has to do with the companies response towards AI. Maybe this will dramatically chague in the near future.
I hope so
Hi! How do you connect the data resourse with chatGPT?
Hello Luke, what is your opinion on using AI in Data science jobs from business perspective and most analysis has to be done based on historical data of customers or any companies data then there is a chance that data we are feeding to AI can be risky. While using chatgpt, what are do's and don'ts especially people working in companies.
Code interpreter is available now. And I've seen some amazing results. Can you do a video to cover that? Tks.
making it as I type this!
It could be fabricating the data because, 1: it couldn't read it properly. 2: it ran out of tokens and couldn't process that data. Those are my two theories.
Hey Luke, do you use big query?
Yeah
I think A.I. should should be used just as another tool for data science and analytics. It should just make things easier.
So the biggest question of all, will it make data analysts less and less in demand in the future?
Hi Luke! Your chat GPT is plus?
Yes
Such an awesome video, I am just getting into data analytics as a career and am worried if Im wasting my time as these AI tools may well do the task 10x better in a millisecond of the time but it seems that's not about to happen soon.
In your video between seconds 7 and 13 you linked a spreadsheet to chatpgpt. It went by in a flash, how do you actually step by step link them?
check out my videos after this one; I go into more detail on this in both of them
@@LukeBarousse thank you Luke
4 out of 5 is an A+? Confirmed Luke plays JRPGs and knows S tier is the highest.
One of the frustrating things I hear about are those companies that have blocked these AI tools from being used because they deal with sensitive data and employees can’t be trusted to use it properly.
Haha you know it’s s-tier!
And yeah companies have to start thinking of what there workaround is. Just saw the other day Google is banning use of Bard for Google employees to input their data into. So craxy
What do you think about the Code Interpreter in ChatGPT (Able to generate charts, maps, data viz, clean datasets)? There're tweets and news explaining that "anyone can be a data analyst now" or "anyone can be hired a data analyst for only $20/month". I've been studying for 3 months and increasingly afraid that in the future data analysts are not needed, am I too late to become a data analyst?
Not at all late. I feel code interpreter is very complimentary and that ai tools in general will be a common addition to data analyst tools.
@@LukeBarousse Awesome thanks! I'll keep on studying and rocking the data world
What plugin are you using to get ChatGPT to look at data?
Noteable
Does connecting the excel to chatgpt does that make your data less secure? Or does it expose the data / share it in anyway?
This is an excellent video. Cudos!
Can you tell us how to use these plugins to make these analyses and visualizations directly in ChatGPT through plain text prompts? This would be the push I needed to move from the free to paid ChatGPT Plus account.
My follow on videos go into this a lot more
How accurate do you think ChatGPT's analyses are? Do you compare your results with theirs to estimate a percentage of error?
Great question. It really depends on the prompting IMO. If I keep it on track with descriptive prompts I feel it performs much better and don’t really have to worry about hallucinations.
cool video thank you
Do you have any concerns about sharing confidential data with chatgpt, since you are using it for work?
For me, no. The data I have I don't have security concerns with. And I think many fall into this category; But others that may have confidential data (HIPPA data, etc.); yeah they def shouldn't put that in here
@@LukeBarousse great, thanks! Do you think anonymizing HIPPA data is sufficient? When I said share confidential data, I meant more like a company's internal proprietary data. Everyone I've asked will side with caution and say "Be careful what you share with ChatGPT, if you wouldn't want the data published on the internet then don't share it" ...but then I don't see how we can unlock it's full potential. Do you have any additional thoughts?
How graphs appeared on chatgpt?
Which plug-in is he using for searching the best up to date information on the internet?
I was using their Bing Plug-in.. they also have different plug-ins for searching for scientific research papers and articles... but mainly I use Bing
@@LukeBarousse thanks for the response. Appreciate your work.
I was wondering if someone could help me plan out a “Roadmap to first Data Analytics Job”. I’ve seen so many different videos, each of them recommending different paths. Does someone have a preferred “path” to take for someone starting with little to no data experience?
🎯 Key Takeaways for quick navigation:
00:00 📊 *Using ChatGPT for Data Analysis*
- Demonstrates using ChatGPT for data analysis tasks, significantly reducing time compared to manual methods.
- Utilizes ChatGPT to connect to a spreadsheet, analyze job postings, and generate visualizations like bar graphs and salary calculations.
01:23 🤖 *Understanding AI's Impact on Data Jobs*
- Explores the impact of AI on both individuals and businesses within the realm of data jobs.
- Outlines a two-step approach involving data research and analysis using ChatGPT to investigate sentiments and effects of AI on data science roles.
02:31 💼 *Exploring ChatGPT Capabilities*
- Provides insights into ChatGPT, an AI model by OpenAI, highlighting its effectiveness in data analysis.
- Discusses the benefits of the pro version, including access to plugins for enhanced functionality.
04:05 🔍 *Searching for Information on AI Impact*
- Discusses limitations of ChatGPT in accessing real-time data and the use of plugins like Bing for internet searches.
- Utilizes ChatGPT to analyze PDF documents and surveys, extracting insights about AI's impact on data-related tasks.
05:55 📊 *Analyzing Developer Sentiment Towards AI*
- Examines survey results from developers regarding their sentiment towards AI in the workplace.
- Utilizes ChatGPT to summarize and analyze survey findings, revealing positive sentiments among data professionals towards AI technology.
08:01 💼 *Assessing Business Perspectives on AI*
- Explores how businesses perceive AI's impact on economic output and workforce dynamics.
- Discusses potential challenges such as privacy concerns while emphasizing the benefits of AI adoption for companies.
09:10 📊 *Analyzing Job Postings for AI Skills*
- Utilizes ChatGPT and plugins for data analysis, focusing on job postings to understand the demand for AI skills.
- Faces challenges such as dataset size limitations and sampling issues but generates insights into the prevalence of AI-related skills in job descriptions.
11:45 🔍 *Exploring AI Tools in Job Descriptions*
- Investigates the inclusion of AI-related tools in job descriptions, revealing low percentages of job postings requiring AI skills.
- Analyzes salary differentials among job titles, highlighting discrepancies in compensation for roles requiring AI expertise.
13:09 💡 *Concluding Remarks on AI Implementation*
- Summarizes findings indicating a disparity between individual interest in AI and its adoption in job postings.
- Calls for further exploration into the reasons behind this disconnect and suggests future research directions.
Made with HARPA AI
can i become data analyst by using phone or even learn by mobile
Conclusion on hiring did not age well. ML and Gen AI skills hiring is defining the tech economy in 2024 and probably through 2027, at least, though of course there will be biggish layoffs, as well.
Incredible workflows though, thank you!
Conclusion of the video: Data Analysts want to use AI tools, but the industry just want its data analysis to be done, anyhow.
Is the sentiment still optimistic after code interpreter release
Yeah. From what I’ve found Code interpreter is faster and has better insights. So I think we are in a good direction
Schwarzenegger 🎉
🙌🏼🙌🏼🙌🏼
Nice green screen work 👌
the hallucinations are such a bummer. hard to have confidence when it routinely makes stuff up. qc nightmare. i know it will get better, but for business use cases it's still pretty sketch at the moment. very much swim at your own risk.
Garmin Fenix 7X?
Small computer? It looks pretty hefty to me!
It’s small compared to your hefty M1 Mac!!
What is the installed plugin?
Which one?
great content!
Code interpreter: hold my programming
Companies not interested or HR drones not interested?
Pretty cool!
“How I use chat GPT as a Data Analyst” or “How I got fired for feeding company’s data to public domain”.
Will ChatGPT ever have internet access and be kept up to date?
I am feeling detracted for using chatgpt for work because I might be using private data. How do you guys deal with this?
Mask the sensitive data. If I use it for work, I change names of things with the name of the client to the word "client" for example, if there is one client. If there are personal ID numbers, just split them and concatenate them with a text string of something unique for that row of data, that is not sensitive data, so you can revert it easily.
Still waiting for that Chat GPT to be able to apply sunblock
ChatGpt can replace Data analytics job ???
Companies are offering less wage for job with AI tools requirements because they think it will be less difficult.