Thanks Chris. This is great. I imagine these tools will continue evolving to less and less guidance to do it by itself, except those items where human expertise and context is needed.
Hey, Chris. Really nice demonstration! I appreciated you including the QA issues in the data. It would be interesting to see if ADA could handle some more advanced techniques for handling things like missing data, such as multiple imputation. I've tested ChatGPT on the programming side and found the same results as you, it needs a pretty heavy hand from a skilled analyst to create the correct code. Another aspect to consider is that many tests of AI and coding/analysis are focusing on a limited number of analytic steps. Your test here is probably the most involved I have seen since you tackled validation, correction, and visualization. But most analyses are substantially more involved in terms of the scope and scale of the analysis. This is another place where the human expertise, and ability to think 10 steps ahead, are critical. Setting up a dataset/analysis to handle multiple questions, provide more extensive insights, and incorporate contextual knowledge of the business problem environment is still something that AI will have problems with.
Sir this a remarkable explore of how accurate can be ChatGPT AI if guided appropriately. Please make a project of book lending service and return reports of a book library and so on so forth.
Very cool video! My take is that while ChatGPT may be helpful in terms of analyzing the data, there is still a need for a skilled analyst to guide where the analysis should go. I agree with you that ChatGPT can work as a very awesome sidekick and not the real superhero haha
Hi Chris! Really nice demonstration! As analysts, we are all kind of thoughtful that all these technologies may soon take our jobs down. However, your video shows that there is still a need for someone to trigger for analysis. Hope we can collaborate with the AI fairly.
Thanks Chris. This is great. I imagine these tools will continue evolving to less and less guidance to do it by itself, except those items where human expertise and context is needed.
Hey, Chris. Really nice demonstration! I appreciated you including the QA issues in the data. It would be interesting to see if ADA could handle some more advanced techniques for handling things like missing data, such as multiple imputation. I've tested ChatGPT on the programming side and found the same results as you, it needs a pretty heavy hand from a skilled analyst to create the correct code. Another aspect to consider is that many tests of AI and coding/analysis are focusing on a limited number of analytic steps. Your test here is probably the most involved I have seen since you tackled validation, correction, and visualization. But most analyses are substantially more involved in terms of the scope and scale of the analysis. This is another place where the human expertise, and ability to think 10 steps ahead, are critical. Setting up a dataset/analysis to handle multiple questions, provide more extensive insights, and incorporate contextual knowledge of the business problem environment is still something that AI will have problems with.
Sir this a remarkable explore of how accurate can be ChatGPT AI if guided appropriately. Please make a project of book lending service and return reports of a book library and so on so forth.
Very cool video! My take is that while ChatGPT may be helpful in terms of analyzing the data, there is still a need for a skilled analyst to guide where the analysis should go. I agree with you that ChatGPT can work as a very awesome sidekick and not the real superhero haha
Hi Chris! Really nice demonstration! As analysts, we are all kind of thoughtful that all these technologies may soon take our jobs down. However, your video shows that there is still a need for someone to trigger for analysis. Hope we can collaborate with the AI fairly.
Outstanding overview of key best practices! Thank you very much for this! :)
Thanks Chris!