What's new
By:

The Defense Thread

NAOS_pt2 said:
Here's a super fun conversation I just had with chatGPT (edit down to avoid tl;dr). Enjoy!!!

NAOS:
Can you please share an outline of a collaborative workflow that involves the sharing of data from a data exploration exercise? To be more specific, I'm interested in how teams work together starting with the early/provisional results of a data exploration. What kind of information is typically acceptable to share? What are the best practices for the group that receives this information?

ChatGPT:
Certainly! Collaborative data exploration is a crucial step in any data-driven project. Here's an outline of a collaborative workflow that involves the sharing of data from a data exploration exercise, including what to share and best practices for the teams involved:

1. Data Exploration Phase:
a. Data Collection: …
b. Data Cleaning: …
c. Data Normalization and Scaling: …:
d. Data Analysis: Conduct initial data analysis…

2. Sharing Initial Findings:
- Summarize the data exploration process and the context of the data.
- Highlight key findings, patterns, and trends discovered.
...

NAOS:
Other than data normalization and data scaling issues, is it generally good practice to avoid any additional data transformations until after you've looked at the data you've received?

ChatGPT:
In most cases, it's a good practice to avoid making additional data transformations until after you've had a chance to thoroughly examine the data you've received during the initial data exploration phase. Data transformations, such as feature engineering or complex scaling methods, can introduce unintended biases or distort the underlying patterns in the data. Here are some reasons to postpone extensive data transformations until after the initial data inspection:

Preserve Data Integrity: …

Assess Data Structure: …

Identify Data Issues: …

Evaluate Transformations' Necessity: …

Avoid Overfitting: …

Iterative Approach:...
...

NAOS:
Is it fair to tell your interlocutors that you aren't presenting any "conclusions" at the end of a data exploration, but instead merely showing them "what initially pops out", or something like that?

ChatGPT:
Yes, it is fair and advisable to communicate to your stakeholders or interlocutors that you are not presenting "conclusions" but rather sharing what initially stands out or what you've observed in the data exploration process. This kind of transparency and clear communication is important for several reasons: …
Click to expand...
And I asked chatgpt for an ice cream recipe and it called for 1/4 cup of kosher salt. Yum!
 
Agent#00 said:
That still doesn't change the fact that there are 49million kangaroos in Australia and 3.5million people in Uruguay which means if the kangaroos were to invade Uruguay, each person will have to fight 14 kangaroos ..
Click to expand...
That's terrifying. If they invaded America it would just be known as the NRA vindication invasion.
 
Bruh this is hilarious. For one, NAOS can't get enough of me even after the supposed "mute" haha. Second, he's asking very pointed questions to freaking ChatGPT to try to look smart haha. This opened up a world of hilarity I did not even know what was possible.

Ask ChatGPT the questions I stated. Even just the first one, "Is the pearson coefficient a data transformation?". He doesn't know what a Pearson coefficient is and I proved that by letting him ramble on about how it's a data transformation. It is not. Seriously, it takes 10 seconds to as ChatGPT to do this. Anyone can do this and everybody including NAOS can see that he doesn't know what he's talking about.

This is what happens when you have someone larping as a data literate person. It's ok if you're not, there was plenty of discussion without the Pearson coefficient. My argument stands with or without it. But it is freaking hilarious that NAOS does not believe he has to provide an explanation (while also demanding explanations from me) because of this strict and rigid process.

What we're seeing is what would happen if someone tried to apply for a data analysis related job without any experience and only the help of Chatgpt hahahaha. He tried so hard to make Chatgpt say it's inappropriate to use Pearson during "data exploration". And yet it still spit out this answer:

In addition to the Pearson Correlation Coefficient, data analysts often use a variety of statistical and visualization methods during data exploration.
Click to expand...

Hahahahahaha. So what is it...Can we, or can we not use Pearson during data exploration? Is it, or is it not a data transformation?
 
I have multiple degrees related to analytics. I've worked in analytics for years. I'm literally at an analytics conference right now where I presented. I truly hate to be the guy who appeals to authority. Having that experience does not make me right or wrong, I am not playing this card to say "I'm right because of expertise or authority". I feel like I'm right because of the rationale I provided and no expertise is needed to get to the same points and conclusions.

I'm just mentioning that because it's freaking hilarious for someone to try to lecture me on what the correct process is for data exploration. This "data exploration" angle is being used as the justification as to why NAOS does not have to have reasoning or rationale and it also the reason why I cannot use my own rationale after being attacked for providing no rationale.

Asking pointed at ChatGPT to try to defend this stance only to have it spit out the wrong thing. Funniest thing ever. Pearson is not data transformation and it can also be used as a part of data exploration, all provided in the response to NAOS's prompts.
 
NAOS:
I've found that it can be easy to miscommunicate about what constitutes a data transformation.

chatGPT:
Absolutely, the term "data transformation" can be interpreted in various ways, leading to potential miscommunication. To avoid confusion, it can be helpful to clarify the context and provide specific examples when discussing data transformations....

NAOS:
I was taught that the term "data transformation" applied to any practice that took the original dataset and added or changed values in order to investigate relationships.

chatGPT:
The term "data transformation" can indeed have a broad or specific interpretation depending on the context and the field of study. Your understanding aligns with a more inclusive definition where any practice altering or adding values for investigating relationships is considered a data transformation.
 
KqWIN said:
Bruh this is hilarious. For one, NAOS can't get enough of me even after the supposed "mute" haha. Second, he's asking very pointed questions to freaking ChatGPT to try to look smart haha. This opened up a world of hilarity I did not even know what was possible.

Ask ChatGPT the questions I stated. Even just the first one, "Is the pearson coefficient a data transformation?". He doesn't know what a Pearson coefficient is and I proved that by letting him ramble on about how it's a data transformation. It is not. Seriously, it takes 10 seconds to as ChatGPT to do this. Anyone can do this and everybody including NAOS can see that he doesn't know what he's talking about.

This is what happens when you have someone larping as a data literate person. It's ok if you're not, there was plenty of discussion without the Pearson coefficient. My argument stands with or without it. But it is freaking hilarious that NAOS does not believe he has to provide an explanation (while also demanding explanations from me) because of this strict and rigid process.

What we're seeing is what would happen if someone tried to apply for a data analysis related job without any experience and only the help of Chatgpt hahahaha. He tried so hard to make Chatgpt say it's inappropriate to use Pearson during "data exploration". And yet it still spit out this answer:



Hahahahahaha. So what is it...Can we, or can we not use Pearson during data exploration? Is it, or is it not a data transformation?
Click to expand...
yeah, I decided to unmute this when I had a few free minutes at the office today. I only go into the office once a month, and I knew there was a dude here who would laugh his *** off at your meltdown. It was right.
 
KqWIN said:
I have multiple degrees related to analytics. I've worked in analytics for years. I'm literally at an analytics conference right now where I presented. I truly hate to be the guy who appeals to authority. Having that experience does not make me right or wrong, I am not playing this card to say "I'm right because of expertise or authority". I feel like I'm right because of the rationale I provided and no expertise is needed to get to the same points and conclusions.

I'm just mentioning that because it's freaking hilarious for someone to try to lecture me on what the correct process is for data exploration. This "data exploration" angle is being used as the justification as to why NAOS does not have to have reasoning or rationale and it also the reason why I cannot use my own rationale after being attacked for providing no rationale.

Asking pointed at ChatGPT to try to defend this stance only to have it spit out the wrong thing. Funniest thing ever. Pearson is not data transformation and it can also be used as a part of data exploration, all provided in the response to NAOS's prompts.
Click to expand...
There isn't a correct process for data exploration. You're the person who's policing what a data exploration is and isn't. Not me. How can you be twisted about this feature of our back-n-forth? Answer: cuz you're melting down and lashing out... got all tied up in a knot.

The chatGPT stuff is just for fun. I went with that angle because you can't have a conversation, apparently.
 
Last edited:
NAOS_pt2 said:
There isn't a correct process for data exploration.

The chatGPT stuff is just for fun. I went with that angle because you can't have a conversation, apparently.
Click to expand...

Yeah, I know that. You can't say this and also say that it must be done in such a specific way lmao. I was not the one saying you cannot use the Pearson coefficient in data exploration because it's a data transformation.

The ChatGPT stuff is a lot of fun. Even your massaged prompts explicitly justify my use of the Pearson coefficient.
 
You must log in or register to reply here.

Share this page
Back
Top