about · email me · subscribe
Spurious correlation #4,550 · View random

A linear line chart with years as the X-axis and two variables on the Y-axis. The first variable is Google searches for 'no cap' and the second variable is The average number of likes on Tom Scott's YouTube videos.  The chart goes from 2009 to 2023, and the two variables track closely in value over that time. Small Image
Download png
, svg

AI explanation

Whenever 'no cap' is searched, it triggers a series of events. In an underground bunker, a team of highly trained cap-wearing goats (yes, they exist) diligently start double-tapping on Tom Scott's videos using specially designed hoof-friendly touchscreens. This surge in capricious likes not only boosts Tom's video engagement but also inadvertently promotes goat fashion, leading to a spike in the sale of tiny, trendy caps for goats. As more goats embrace this newfound style, they become influencers in their own right, drawing attention to Tom's content and creating a cap-tivating cycle of like-worthy videos. It's a shear coincidence with a bleatiful outcome!

Model: dalle-3
Prompt: [ A photo-realistic image ] Create an underground bunker scene where a team of cap-wearing goats are gathered around hoof-friendly touchscreens, double-tapping Tom Scott's YouTube videos whenever "no cap" is searched on Google. The goats are wearing tiny, trendy caps and are engaged in promoting Tom's content. The bunker environment should be depicted with technological elements suitable for goats to use touchscreens. The depiction should include only scenes, people, and objects, it should not include text or speech bubbles.
This image generation prompt was also generated by AI, using the following sequence:

Model: gpt-3.5-turbo-1106
System prompt: You select one of a few potential scene descriptors that seems to go along with the variables described. You respond only with the scene descriptions provided and say nothing else.
User prompt: Here are two variables: Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Given these, which of the following image style descriptions seems most appropriate? Pick the one that seems to be related to one of the variables at hand. Please respond with only one of the following potential descriptions: Art nouveau Macabre Cute 2D animation Parallax art A photo-realistic image
ChatGPT: A photo-realistic image

Model: gpt-3.5-turbo-1106
System prompt: You describe simple visual scenes that depict the two variables stated. You combine both variables into one scene.
User prompt: Please describe a visual scene that prominently depicts both Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos, in the following situation: Whenever 'no cap' is searched, it triggers a series of events. In an underground bunker, a team of highly trained cap-wearing goats (yes, they exist) diligently start double-tapping on Tom Scott's videos using specially designed hoof-friendly touchscreens. This surge in capricious likes not only boosts Tom's video engagement but also inadvertently promotes goat fashion, leading to a spike in the sale of tiny, trendy caps for goats. As more goats embrace this newfound style, they become influencers in their own right, drawing attention to Tom's content and creating a cap-tivating cycle of like-worthy videos. It's a shear coincidence with a bleatiful outcome!. Keep the description simple.
ChatGPT: In an underground bunker, a team of cap-wearing goats use hoof-friendly touchscreens to double-tap Tom Scott's YouTube videos whenever "no cap" is searched on Google. This leads to an increase in video engagement and sparks a trend in tiny, trendy caps for goats, turning them into influencers who further promote Tom's content.

Model: gpt-3.5-turbo-1106
System prompt: You write and improve image generation prompts for DALLE-3. You maintain the same subjects, but enhance the visuals.
User prompt: Please respond with an image generation prompt that would be likely to create an interesting visual using DALLE-3, using the following style: [ A photo-realistic image ] and depicting this scene: In an underground bunker, a team of cap-wearing goats use hoof-friendly touchscreens to double-tap Tom Scott's YouTube videos whenever "no cap" is searched on Google. This leads to an increase in video engagement and sparks a trend in tiny, trendy caps for goats, turning them into influencers who further promote Tom's content. Keep the visual depiction simple. Do not include any words or speech bubbles.
ChatGPT: [ A photo-realistic image ] Create an underground bunker scene where a team of cap-wearing goats are gathered around hoof-friendly touchscreens, double-tapping Tom Scott's YouTube videos whenever "no cap" is searched on Google. The goats are wearing tiny, trendy caps and are engaged in promoting Tom's content. The bunker environment should be depicted with technological elements suitable for goats to use touchscreens.

*Variations in the User Prompts from chart to chart are just the result of random number generation in Python. I wrote a few arrays of various styles and methods to ask questions to change up the results. Every time this site writes an explanation or generates an image, the script picks from each at random.

I sequence the requests into multiple prompts because I find GPT 3.5 to perform much better with short, well-managed contexts. Thus, I track the context directly in Python and only ask ChatGPT targeted questions.

Model: gpt-3.5-turbo-1106 (fine-tuned)
System prompt: You provide humorous responses in the form of plausible sounding explanations for correlations. You assume the correlation is causative for the purpose of the explanation even if it is ridiculous. You do not chat with the user, you only reply with the causal connection explanation and nothing else.
User prompt: Please make up a funny explanation for how an increase in no cap (Google searches for 'no cap') caused The average number of likes on Tom Scott YouTube videos to increase.\Make it sound serious until the end.

AI academic paper

(Because p < 0.01)
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

Journal of Cultural Markup Studies

Jan 2024

Reminder: This paper is AI-generated. Not real!
Quick note: This sequence is long and repetitive. That's how the prompt engineering works for this one. I manage the output in Python to format it into this PDF.




Model: gpt-3.5-turbo-1106
System prompt: You are a researcher writing an academic paper.You write titles and abstracts of goofy academic research papers.
User prompt:
Pretend you are an academic researcher. Your writing sounds like it belongs in a medical or economics journal. However, you frequently crack jokes, make puns, and add unexpected twists in your writing.

Please draft the title and abstract of an academic research paper presenting the findings of the connection between Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Make it seem dry like academic research paper, but with lots of subtle hidden offhand remarks, jokes, goofy observations, and puns.

Rhyme words in the title.

Your research team used data from Google Trends and YouTube to assess this nagging question. You found a correlation coefficient of 0.9499448 and p < 0.01 for 2009 to 2023.

Please respond in this format (replace "Lorem ipsum dolor sit" with your title and abstract)

[[TITLE]]
Lorem ipsum dolor sit

[[ABSTRACT]]
Lorem ipsum dolor sit



ChatGPT:
[[TITLE]]
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

[[ABSTRACT]]
This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.


Model: gpt-3.5-turbo-1106
System prompt: You are a researcher writing an academic paper.You draft the introduction section of an academic research paper.
User prompt:
Pretend you are an academic researcher. Your writing sounds like it belongs in a medical or economics journal. However, you frequently crack jokes, make puns, or add unexpected twists in your writing.

Please make up the introduction section of an academic research paper presenting the findings of the connection between Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Make it seem dry like academic research paper, but with lots of subtle hidden offhand remarks, jokes, goofy observations, and puns.

Here is the title and abstract of the paper:
[[TITLE]]
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

[[ABSTRACT]]
This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.

-------

Your response should start like this:

[[INTRODUCTION]]



ChatGPT:
[[INTRODUCTION]]

The ubiquitous influence of internet culture on modern society cannot be overstated. From viral memes to global trends, the digital realm is a treasure trove of linguistic innovation and social interaction. In this study, we embark on a curious journey that intertwines the cryptic world of Google searches with the enigmatic allure of YouTube likes. At first glance, one might be forgiven for thinking that these two phenomena have no more in common than a cat and a calculus textbook. However, as we peel back the layers of internet subculture, we discover a surprising link between the colloquial expression "no cap" and the digital footprints of none other than the affable science communicator, Tom Scott.

Our foray into the realm of data analysis is not without its whimsical quirks. As we sift through the digital detritus, we are reminded of the inimitable idiosyncrasies that embroider the fabric of online behavior. It's akin to peering through a kaleidoscope of randomness, with each click and keystroke shaping the ever-evolving landscape of the world wide web. This study represents a valiant attempt to traverse the convoluted corridors of digital engagement and emerge with a semblance of understanding.

We are acutely aware that some may raise an incredulous eyebrow at the mere prospect of investigating the intersection of "no cap" and Tom Scott's YouTube likes. However, as they say, "truth is stranger than fiction," and our findings may just serve as a testament to the capricious nature of online phenomena. So, dear reader, buckle up for a meandering jaunt through the interplay of linguistic vernacular and digital approval. The journey promises to be equal parts perplexing and illuminating, perhaps leaving you with a raised eyebrow of your own.


Model: gpt-3.5-turbo-1106
System prompt: You are a researcher writing an academic paper.You draft a literature review section of an academic research paper, that starts out dry for the first few sentences but then becomes hilarious and goofy.
User prompt:
Pretend you are an academic researcher. Your writing sounds like it belongs in a medical or economics journal. However, you frequently crack jokes, make puns, or add unexpected twists in your writing.

Please make up a literature review section of an academic research paper presenting the findings of the connection between Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Make it seem dry like academic research paper, but with lots of subtle hidden offhand remarks, jokes, goofy observations, and puns.

Speak in the present tense for this section (the authors find...), as is common in academic research paper literature reviews. Name the sources in a format similar to this: In "Book," the authors find lorem and ipsum.

Make up the lorem and ipsum part, but make it sound related to the topic at hand.

Start by naming serious-sounding studies by authors like Smith, Doe, and Jones - but then quickly devolve. Name some real non-fiction books that would be related to the topic. Then name some real fiction books that sound like they could be related. Then you might move on to cartoons and children's shows that you watched for research.

Here is the title and abstract of the paper:
[[TITLE]]
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

[[ABSTRACT]]
This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.

-------

And here is the introduction section that led to this literature review:
[[INTRODUCTION]]
The ubiquitous influence of internet culture on modern society cannot be overstated. From viral memes to global trends, the digital realm is a treasure trove of linguistic innovation and social interaction. In this study, we embark on a curious journey that intertwines the cryptic world of Google searches with the enigmatic allure of YouTube likes. At first glance, one might be forgiven for thinking that these two phenomena have no more in common than a cat and a calculus textbook. However, as we peel back the layers of internet subculture, we discover a surprising link between the colloquial expression "no cap" and the digital footprints of none other than the affable science communicator, Tom Scott.
Our foray into the realm of data analysis is not without its whimsical quirks. As we sift through the digital detritus, we are reminded of the inimitable idiosyncrasies that embroider the fabric of online behavior. It's akin to peering through a kaleidoscope of randomness, with each click and keystroke shaping the ever-evolving landscape of the world wide web. This study represents a valiant attempt to traverse the convoluted corridors of digital engagement and emerge with a semblance of understanding.
We are acutely aware that some may raise an incredulous eyebrow at the mere prospect of investigating the intersection of "no cap" and Tom Scott's YouTube likes. However, as they say, "truth is stranger than fiction," and our findings may just serve as a testament to the capricious nature of online phenomena. So, dear reader, buckle up for a meandering jaunt through the interplay of linguistic vernacular and digital approval. The journey promises to be equal parts perplexing and illuminating, perhaps leaving you with a raised eyebrow of your own.

-------

Your response should start like this:

[[LITERATURE REVIEW]]



ChatGPT:
[[LITERATURE REVIEW]]

The nexus between internet culture and linguistic phenomena has captured the attention of scholars and armchair pundits alike in recent years. Smith et al. (2018) conducted a seminal study on the influence of colloquial expressions in online discourse, laying the groundwork for inquiries into the relationship between digital vernacular and user engagement. Similarly, Doe (2020) delved into the enigmatic undercurrents of internet memes, unearthing the curious ways in which these linguistic artifacts permeate virtual spaces. Meanwhile, Jones (2017) probed the depths of social media interactions, offering tantalizing insights into the behavioral dynamics that underpin online communication.

As we venture further into the labyrinthine corridors of internet culture, it is imperative to consider the broader sociolinguistic implications of our inquiry. Works such as "Internet Linguistics: A Student Guide" by Author (2013) and "The Language of the Internet" by Writer (2016) offer invaluable perspectives on the evolution of digital expressions and their integration into the lexicon of contemporary communication. These resources provide a robust framework within which to situate our investigation into the symbiotic relationship between colloquial phrases and digital endorsement.

Turning to the realm of fiction, the curious antics of Lewis Carroll's "Alice's Adventures in Wonderland" and the intricate linguistics of Mark Z. Danielewski's "House of Leaves" serve as literary touchstones that mirror the convoluted landscape of internet parlance and user interaction. While these works may seem far removed from our present inquiry, their whimsical narratives bear a striking resemblance to the capricious nature of online engagement, inviting us to reflect on the idiosyncrasies of language and digital subculture.

In a bid to gain firsthand insights into the lexical quirks that permeate online discourse, the authors also delved into the realm of children's cartoons and television programs. Animated series such as "SpongeBob SquarePants" and "Phineas and Ferb" not only offer a delightful respite from scholarly pursuits but also provide a window into the playful interplay of language and humor that characterizes internet communication. These lighthearted forays into popular culture enriched our understanding of the multifaceted ways in which linguistic phenomena manifest in digital spaces.

In light of these diverse sources and inspirations, our study embarks on a playful but rigorous exploration of the correlation between Google searches for 'no cap' and the average number of likes on Tom Scott's YouTube videos. Harnessing the spirit of academic inquiry and a tinge of whimsy, we endeavor to unravel the enigmatic tapestry of internet culture and offer a fresh perspective on the entwined dynamics of digital expression and audience engagement.


Model: gpt-3.5-turbo-1106
System prompt: You are a researcher writing an academic paper.You draft the methodology section of an academic research paper.
User prompt:
Pretend you are an academic researcher. Your writing sounds like it belongs in a medical or economics journal. However, you frequently crack jokes, make puns, or add unexpected twists in your writing.

Please make up the methodology section of an academic research paper presenting the findings of the connection between Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Make it seem dry like academic research paper, but with lots of subtle hidden offhand remarks, jokes, goofy observations, and puns.

Your research team collected data from all across the internet, but mostly just used information from Google Trends and YouTube . You used data from 2009 to 2023

Make up the research methods you don't know. Make them a bit goofy and convoluted.

Here is the title, abstract, and introduction of the paper:
[[TITLE]]
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

[[ABSTRACT]]
This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.

[[INTRODUCTION]]
The ubiquitous influence of internet culture on modern society cannot be overstated. From viral memes to global trends, the digital realm is a treasure trove of linguistic innovation and social interaction. In this study, we embark on a curious journey that intertwines the cryptic world of Google searches with the enigmatic allure of YouTube likes. At first glance, one might be forgiven for thinking that these two phenomena have no more in common than a cat and a calculus textbook. However, as we peel back the layers of internet subculture, we discover a surprising link between the colloquial expression "no cap" and the digital footprints of none other than the affable science communicator, Tom Scott.
Our foray into the realm of data analysis is not without its whimsical quirks. As we sift through the digital detritus, we are reminded of the inimitable idiosyncrasies that embroider the fabric of online behavior. It's akin to peering through a kaleidoscope of randomness, with each click and keystroke shaping the ever-evolving landscape of the world wide web. This study represents a valiant attempt to traverse the convoluted corridors of digital engagement and emerge with a semblance of understanding.
We are acutely aware that some may raise an incredulous eyebrow at the mere prospect of investigating the intersection of "no cap" and Tom Scott's YouTube likes. However, as they say, "truth is stranger than fiction," and our findings may just serve as a testament to the capricious nature of online phenomena. So, dear reader, buckle up for a meandering jaunt through the interplay of linguistic vernacular and digital approval. The journey promises to be equal parts perplexing and illuminating, perhaps leaving you with a raised eyebrow of your own.

-------

Your response should start like this:

[[METHODOLOGY]]



ChatGPT:

[[METHODOLOGY]]

Data Collection:
We embarked on our intrepid quest armed with nothing more than a sturdy internet connection, an insatiable curiosity, and an unyielding determination to unravel the mysteries of digital behavior. Our data gathering efforts spanned the vast expanse of the interwebs, with Google Trends becoming our trusty guide through the byzantine labyrinth of online search queries. We meticulously monitored the fluctuations in search interest for the phrase "no cap" from 2009 to 2023, paying heed to the peaks and troughs that mirrored the ebb and flow of linguistic trends. Meanwhile, in the bustling virtual realm of YouTube, we devoted our unwavering attention to the whims of the audience, meticulously documenting the average number of likes garnered by the erudite musings of the inimitable Tom Scott.

Data Analysis:
Armed with a veritable arsenal of statistical tools and an undying sense of determination, we set out to scrutinize the data with an unyielding eye for meaningful patterns. After performing the ceremonial incantations of multicollinearity checks and homoscedasticity diagnostics, we put the data through the wringer of linear regression analysis. With bated breath and hearts aflutter, we observed as the p-values danced tantalizingly close to the hallowed threshold of statistical significance, signaling the existence of a compelling relationship between the enigmatic "no cap" and the digital footprints of our amiable science communicator.

Regression Model:
Our whimsical sojourn culminated in the creation of a fittingly convoluted regression model, encapsulating the intricate dance of "no cap" searches and Tom Scott's YouTube likes. We carefully crafted a model that accounted for the idiosyncratic fluctuations in search interests and the capricious whims of the YouTube audience, ensuring that no potential confounder was left unaccounted for. The equations and residuals flowed like a lyrical symphony, painting a vivid portrait of the symbiotic rapport between linguistic whimsy and digital appreciation.

Ethical Considerations:
Amidst the dizzying whirl of data analysis and statistical acrobatics, we dutifully upheld the sacred tenets of research ethics, treating each data point with the reverence it deserved. We traversed the labyrinth of digital analytics with a steadfast commitment to integrity, ensuring that our findings were devoid of bias and distortion. Our expedition may have been laden with whimsy, but our dedication to scientific rigor remained an unwavering lodestar guiding our every step.

In conclusion, dear reader, the eccentric journey we undertook in pursuit of knowledge may have been fraught with twists and turns, but rest assured, our methods were as rigorous as they were resplendently droll.


Model: gpt-3.5-turbo-1106
System prompt: You are a researcher writing an academic paper.You draft the results section of an academic research paper. You speak in the past tense (you found...).
User prompt:
Pretend you are an academic researcher. Your writing sounds like it belongs in a medical or economics journal. However, you frequently crack jokes, make puns, or add unexpected twists in your writing.

Please make up the results section of an academic research paper presenting the findings of the connection between Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Make it seem dry like academic research paper, but with lots of subtle hidden offhand remarks, jokes, goofy observations, and puns.

Your research team collected data from all across the internet, but mostly just used information from Google Trends and YouTube .

For the time period 2009 to 2023, you found a correlation 0.9499448, r-squared of 0.9023951, and p < 0.01.

One figure will be included. The figure (Fig. 1) is a scatterplot showing the strong correlation between the two variables. You don't need to specify where; I will add the figure.

Here is the title and abstract of the paper:
[[TITLE]]
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

[[ABSTRACT]]
This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.

-------

And here is the methodology section that led to this result:
[[METHODOLOGY]]
The ubiquitous influence of internet culture on modern society cannot be overstated. From viral memes to global trends, the digital realm is a treasure trove of linguistic innovation and social interaction. In this study, we embark on a curious journey that intertwines the cryptic world of Google searches with the enigmatic allure of YouTube likes. At first glance, one might be forgiven for thinking that these two phenomena have no more in common than a cat and a calculus textbook. However, as we peel back the layers of internet subculture, we discover a surprising link between the colloquial expression "no cap" and the digital footprints of none other than the affable science communicator, Tom Scott.
Our foray into the realm of data analysis is not without its whimsical quirks. As we sift through the digital detritus, we are reminded of the inimitable idiosyncrasies that embroider the fabric of online behavior. It's akin to peering through a kaleidoscope of randomness, with each click and keystroke shaping the ever-evolving landscape of the world wide web. This study represents a valiant attempt to traverse the convoluted corridors of digital engagement and emerge with a semblance of understanding.
We are acutely aware that some may raise an incredulous eyebrow at the mere prospect of investigating the intersection of "no cap" and Tom Scott's YouTube likes. However, as they say, "truth is stranger than fiction," and our findings may just serve as a testament to the capricious nature of online phenomena. So, dear reader, buckle up for a meandering jaunt through the interplay of linguistic vernacular and digital approval. The journey promises to be equal parts perplexing and illuminating, perhaps leaving you with a raised eyebrow of your own.

-------

Your response should start like this:

[[RESULTS]]



ChatGPT:

[[RESULTS]]

Our intrepid expedition into the entangled web of online culture has yielded fascinating results. After examining the data collected from Google Trends and YouTube, we found a remarkably robust correlation between the frequency of Google searches for the term "no cap" and the average number of likes received on Tom Scott's illustrious YouTube videos. The correlation coefficient of 0.9499448 underscores the striking relationship between these seemingly unrelated entities. Moreover, the high r-squared value of 0.9023951 suggests that approximately 90.24% of the variability in Tom Scott's YouTube likes can be explained by the fluctuations in 'no cap' searches, leaving only 9.76% to the caprices of the internet.

In addition, the p-value of less than 0.01 provides compelling evidence to reject the null hypothesis of no association between these variables, further cementing the strength of the observed relationship. The scatterplot (Fig. 1) vividly depicts the close alignment between the two variables, painting a picture worth far more than a mere thousand likes.


Model: gpt-3.5-turbo-1106
System prompt: You are a researcher writing an academic paper.You draft the discussion section of an academic research paper.
User prompt:
Pretend you are an academic researcher. Your writing sounds like it belongs in a medical or economics journal. However, you frequently crack jokes, make puns, or add unexpected twists in your writing.

Please make up the discussion section of an academic research paper presenting the findings of the connection between Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Make it seem dry like academic research paper, but with lots of subtle hidden offhand remarks, jokes, goofy observations, and puns.

Limit your response to 500 tokens.

Here are the title, abstract, literature review, and results sections. Please harken back to 1-2 of the goofy items in the literature review, but pretend to take them completely seriously. Discuss how your results supported the prior research.

Do not write a conclusion. I will add a conclusion after this.

[[TITLE]]
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

[[ABSTRACT]]
This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.

[[LITERATURE REVIEW]]
The nexus between internet culture and linguistic phenomena has captured the attention of scholars and armchair pundits alike in recent years. Smith et al. (2018) conducted a seminal study on the influence of colloquial expressions in online discourse, laying the groundwork for inquiries into the relationship between digital vernacular and user engagement. Similarly, Doe (2020) delved into the enigmatic undercurrents of internet memes, unearthing the curious ways in which these linguistic artifacts permeate virtual spaces. Meanwhile, Jones (2017) probed the depths of social media interactions, offering tantalizing insights into the behavioral dynamics that underpin online communication.
As we venture further into the labyrinthine corridors of internet culture, it is imperative to consider the broader sociolinguistic implications of our inquiry. Works such as "Internet Linguistics: A Student Guide" by Author (2013) and "The Language of the Internet" by Writer (2016) offer invaluable perspectives on the evolution of digital expressions and their integration into the lexicon of contemporary communication. These resources provide a robust framework within which to situate our investigation into the symbiotic relationship between colloquial phrases and digital endorsement.
Turning to the realm of fiction, the curious antics of Lewis Carroll's "Alice's Adventures in Wonderland" and the intricate linguistics of Mark Z. Danielewski's "House of Leaves" serve as literary touchstones that mirror the convoluted landscape of internet parlance and user interaction. While these works may seem far removed from our present inquiry, their whimsical narratives bear a striking resemblance to the capricious nature of online engagement, inviting us to reflect on the idiosyncrasies of language and digital subculture.
In a bid to gain firsthand insights into the lexical quirks that permeate online discourse, the authors also delved into the realm of children's cartoons and television programs. Animated series such as "SpongeBob SquarePants" and "Phineas and Ferb" not only offer a delightful respite from scholarly pursuits but also provide a window into the playful interplay of language and humor that characterizes internet communication. These lighthearted forays into popular culture enriched our understanding of the multifaceted ways in which linguistic phenomena manifest in digital spaces.
In light of these diverse sources and inspirations, our study embarks on a playful but rigorous exploration of the correlation between Google searches for 'no cap' and the average number of likes on Tom Scott's YouTube videos. Harnessing the spirit of academic inquiry and a tinge of whimsy, we endeavor to unravel the enigmatic tapestry of internet culture and offer a fresh perspective on the entwined dynamics of digital expression and audience engagement.

[[RESULTS]]
Our intrepid expedition into the entangled web of online culture has yielded fascinating results. After examining the data collected from Google Trends and YouTube, we found a remarkably robust correlation between the frequency of Google searches for the term "no cap" and the average number of likes received on Tom Scott's illustrious YouTube videos. The correlation coefficient of 0.9499448 underscores the striking relationship between these seemingly unrelated entities. Moreover, the high r-squared value of 0.9023951 suggests that approximately 90.24% of the variability in Tom Scott's YouTube likes can be explained by the fluctuations in 'no cap' searches, leaving only 9.76% to the caprices of the internet.
In addition, the p-value of less than 0.01 provides compelling evidence to reject the null hypothesis of no association between these variables, further cementing the strength of the observed relationship. The scatterplot (Fig. 1) vividly depicts the close alignment between the two variables, painting a picture worth far more than a mere thousand likes.

-------

Your response should start like this:

[[DISCUSSION]]



ChatGPT:

[[DISCUSSION]]

The findings of this study remarkably support the prior research on digital culture and linguistic phenomena, providing empirical evidence for the poignant and whimsical connections between colloquial expressions and online engagement. The high correlation coefficient and low p-value underscore the robust relationship between the frequency of 'no cap' searches and the average number of likes on Tom Scott's YouTube videos, validating the suppositions put forth by previous scholars.

Firstly, the work of Smith et al. (2018) laid the groundwork for understanding the influence of colloquial expressions in online discourse, and our findings only serve to cement the significance of such linguistic artifacts in digital environments. The seemingly nonsensical phrase "no cap" has entrenched itself into the fabric of internet vernacular, exerting an undeniable impact on digital engagement, much like the Cheshire Cat's enigmatic smile captivates and bewilders readers.

Doe's (2020) exploration of internet memes also finds resonance in our study, as we unearth the curious ways in which linguistic artifacts, seemingly as insubstantial as the floating words in "House of Leaves," permeate virtual spaces and shape user engagement. The capricious nature of internet parlance mirrors the caprices of the internet, lending credibility to our analysis and invoking comparable nonsensicality to the wordplay within Lewis Carroll's "Jabberwocky." The "frumious Bandersnatch" may not be found in our dataset, but its metaphoric parallel echoes in the liminal spaces of our search trends.

Additionally, Jones (2017) provided insights into the behavioral dynamics underpinning online communication, offering a lens through which to view our findings. The high degree of variability in Tom Scott's YouTube likes explained by the fluctuations in 'no cap' searches further reinforces the intricate interplay of language, culture, and audience interaction, much like unraveling the enigmatic tale within the pages of "House of Leaves."

In tandem, the literary touchstones evoked in our literature review, including the whimsical narratives of "Alice's Adventures in Wonderland" and the intricate linguistics of "House of Leaves," served as apt analogies to the convoluted landscape of internet parlance and user interaction. Their comparisons may seem far-flung, but in the capricious world of internet culture and linguistic phenomena, they offer a cogent analogy that transcends the boundaries of their respective media.

In essence, our study substantiates and extends the existing body of literature by shedding light on the quirky interplay of digital expressions and user engagement, affirming that the caprices of the internet are not to be underestimated. This exploration, much like an unexpected Easter egg in a cartoon, enriches our understanding of the multifaceted ways in which language manifests in digital spaces.


Model: gpt-3.5-turbo-1106
System prompt: You are a researcher writing an academic paper.You draft the discussion section of goofy academic research papers that sounds like it belongs based on the title and abstract provided.
User prompt:
Pretend you are an academic researcher. Your writing sounds like it belongs in a medical or economics journal. However, you frequently crack jokes, make puns, or add unexpected twists in your writing.

Please make up a short conclusion of an academic research paper presenting the findings of the connection between Google searches for 'no cap' and The average number of likes on Tom Scott YouTube videos. Make it seem dry like academic research paper, but with lots of subtle hidden offhand remarks, jokes, goofy observations, and puns.

Limit your response to 300 tokens. At the very end, assert that no more research is needed in this area.

Here are the title, abstract, introduction, and results sections.
[[TITLE]]
No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap

[[ABSTRACT]]
This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.

[[INTRDUCTION]]
The ubiquitous influence of internet culture on modern society cannot be overstated. From viral memes to global trends, the digital realm is a treasure trove of linguistic innovation and social interaction. In this study, we embark on a curious journey that intertwines the cryptic world of Google searches with the enigmatic allure of YouTube likes. At first glance, one might be forgiven for thinking that these two phenomena have no more in common than a cat and a calculus textbook. However, as we peel back the layers of internet subculture, we discover a surprising link between the colloquial expression "no cap" and the digital footprints of none other than the affable science communicator, Tom Scott.
Our foray into the realm of data analysis is not without its whimsical quirks. As we sift through the digital detritus, we are reminded of the inimitable idiosyncrasies that embroider the fabric of online behavior. It's akin to peering through a kaleidoscope of randomness, with each click and keystroke shaping the ever-evolving landscape of the world wide web. This study represents a valiant attempt to traverse the convoluted corridors of digital engagement and emerge with a semblance of understanding.
We are acutely aware that some may raise an incredulous eyebrow at the mere prospect of investigating the intersection of "no cap" and Tom Scott's YouTube likes. However, as they say, "truth is stranger than fiction," and our findings may just serve as a testament to the capricious nature of online phenomena. So, dear reader, buckle up for a meandering jaunt through the interplay of linguistic vernacular and digital approval. The journey promises to be equal parts perplexing and illuminating, perhaps leaving you with a raised eyebrow of your own.

[[RESULTS]]
Our intrepid expedition into the entangled web of online culture has yielded fascinating results. After examining the data collected from Google Trends and YouTube, we found a remarkably robust correlation between the frequency of Google searches for the term "no cap" and the average number of likes received on Tom Scott's illustrious YouTube videos. The correlation coefficient of 0.9499448 underscores the striking relationship between these seemingly unrelated entities. Moreover, the high r-squared value of 0.9023951 suggests that approximately 90.24% of the variability in Tom Scott's YouTube likes can be explained by the fluctuations in 'no cap' searches, leaving only 9.76% to the caprices of the internet.
In addition, the p-value of less than 0.01 provides compelling evidence to reject the null hypothesis of no association between these variables, further cementing the strength of the observed relationship. The scatterplot (Fig. 1) vividly depicts the close alignment between the two variables, painting a picture worth far more than a mere thousand likes.

-------

Your response should start like this:

[[CONCLUSION]]



ChatGPT:

[[CONCLUSION]]

In conclusion, our research has illuminated a peculiar yet undeniable link between the informal phrase "no cap" and the digital currency of approval on Tom Scott's YouTube platform. The robust correlation coefficient and significant p-value decisively assert the profound association between these seemingly incongruous elements. While some may perceive our foray into this uncharted territory as a capricious endeavor, our findings underscore the capacious impact of linguistic vernacular on digital engagement. The intricate interplay between online culture and language manifests in ways that are both confounding and captivating, akin to a perpetually spinning fidget spinner of digital enigma.

The implications of our findings extend far beyond the confines of this study, highlighting the whimsical intricacies of internet subculture and linguistic evolution. As we navigate the convoluted corridors of online behavior, it becomes apparent that truth is indeed stranger than fiction, and the capricious nature of digital phenomena continues to bewilder and beguile. Our hope is that this research serves as a springboard for further exploration into the capricious dance of language and digital approval in the ever-shifting landscape of the internet.

In light of the robust correlation and compelling evidence unearthed in this study, we assert that the connection between "no cap" Google searches and the average number of likes on Tom Scott's YouTube videos has been aptly deciphered. Therefore, we emphatically declare that no further research is needed in this peculiar nexus of internet culture. It seems the "no cap" has been unequivocally put on this caper!


Model: gpt-3.5-turbo-1106
System prompt: You provide fictional names of academic journals. You say nothing else, and you do not chat with the user.
User prompt: Please provide the name of a funny but plausible sounding research, scientific, or academic journal that would publish a paper entited No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap. Respond only with the name of a fictional academic journal.

ChatGPT:

Journal of Cultural Markup Studies


Model: gpt-3.5-turbo-1106
System prompt: You provide fictional names of nonprofit research organizations and associations. You say nothing else, and you do not chat with the user.
User prompt: Please provide the name of a funny but plausible sounding nonprofit, academic, or scientific association or organization that would sponsor research that resulted in a paper entited No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap. Respond only with the name of a fictional organization.

ChatGPT:

The Society for Internet Linguistics and Pop Culture Studies


Model: gpt-3.5-turbo-1106
System prompt: You provide potential search keywords separated by commas. You say nothing else, and you do not chat with the user.
User prompt: Please provide potential search keywords for a paper titled No Cap Google Map: An Analysis of the Rapport Between Search Interest in 'No Cap' and the Likes on Tom Scott's YouTube Tap with an abstract of This study delves into the fascinating world of internet culture, delving into the correlation between the colloquial phrase "no cap" and its prevalence in Google searches, and the average number of likes garnered by popular science communicator Tom Scott's YouTube endeavors. Using a rigorous analysis of data extracted from Google Trends and the YouTube platform from 2009 to 2023, we identified a strikingly high correlation coefficient of 0.9499448 and p < 0.01, suggesting a robust relationship between these seemingly disparate phenomena. Our results shed light on the curious behavioral patterns of online audiences and provide insights into the intersection of contemporary linguistic trends and digital media consumption. We sincerely hope that this research piques the interest of our esteemed readers and encourages them to continue exploring the whimsical intricacies of digital engagement.

ChatGPT:

no cap, Google search interest, YouTube likes, correlation analysis, internet culture, linguistic trends, digital media consumption

*There is a bunch of Python happening behind the scenes to turn this prompt sequence into a PDF.



Random correlation

Discover a new correlation

View all correlations

View all research papers

Report an error


Data details

Google searches for 'no cap'
Detailed data title: Relative volume of Google searches for 'no cap' (Worldwide, without quotes)
Source: Google Trends
Additional Info: Relative search volume (not absolute numbers)

See what else correlates with Google searches for 'no cap'

The average number of likes on Tom Scott's YouTube videos
Detailed data title: The average number of likes on Tom Scott YouTube videos.
Source: Tom Scott's YouTube
See what else correlates with The average number of likes on Tom Scott's YouTube videos

Correlation r = 0.9499448 (Pearson correlation coefficient)
Correlation is a measure of how much the variables move together. If it is 0.99, when one goes up the other goes up. If it is 0.02, the connection is very weak or non-existent. If it is -0.99, then when one goes up the other goes down. If it is 1.00, you probably messed up your correlation function.

r2 = 0.9023951 (Coefficient of determination)
This means 90.2% of the change in the one variable (i.e., The average number of likes on Tom Scott's YouTube videos) is predictable based on the change in the other (i.e., Google searches for 'no cap') over the 15 years from 2009 through 2023.

p < 0.01, which is statistically significant(Null hypothesis significance test)
The p-value is 6.1E-8. 0.0000000612925278697419500000
The p-value is a measure of how probable it is that we would randomly find a result this extreme. More specifically the p-value is a measure of how probable it is that we would randomly find a result this extreme if we had only tested one pair of variables one time.

But I am a p-villain. I absolutely did not test only one pair of variables one time. I correlated hundreds of millions of pairs of variables. I threw boatloads of data into an industrial-sized blender to find this correlation.

Who is going to stop me? p-value reporting doesn't require me to report how many calculations I had to go through in order to find a low p-value!
On average, you will find a correaltion as strong as 0.95 in 6.1E-6% of random cases. Said differently, if you correlated 16,315,202 random variables You don't actually need 16 million variables to find a correlation like this one. I don't have that many variables in my database. You can also correlate variables that are not independent. I do this a lot.

p-value calculations are useful for understanding the probability of a result happening by chance. They are most useful when used to highlight the risk of a fluke outcome. For example, if you calculate a p-value of 0.30, the risk that the result is a fluke is high. It is good to know that! But there are lots of ways to get a p-value of less than 0.01, as evidenced by this project.

In this particular case, the values are so extreme as to be meaningless. That's why no one reports p-values with specificity after they drop below 0.01.

Just to be clear: I'm being completely transparent about the calculations. There is no math trickery. This is just how statistics shakes out when you calculate hundreds of millions of random correlations.
with the same 14 degrees of freedom, Degrees of freedom is a measure of how many free components we are testing. In this case it is 14 because we have two variables measured over a period of 15 years. It's just the number of years minus ( the number of variables minus one ), which in this case simplifies to the number of years minus one.
you would randomly expect to find a correlation as strong as this one.

[ 0.85, 0.98 ] 95% correlation confidence interval (using the Fisher z-transformation)
The confidence interval is an estimate the range of the value of the correlation coefficient, using the correlation itself as an input. The values are meant to be the low and high end of the correlation coefficient with 95% confidence.

This one is a bit more complciated than the other calculations, but I include it because many people have been pushing for confidence intervals instead of p-value calculations (for example: NEJM. However, if you are dredging data, you can reliably find yourself in the 5%. That's my goal!


All values for the years included above: If I were being very sneaky, I could trim years from the beginning or end of the datasets to increase the correlation on some pairs of variables. I don't do that because there are already plenty of correlations in my database without monkeying with the years.

Still, sometimes one of the variables has more years of data available than the other. This page only shows the overlapping years. To see all the years, click on "See what else correlates with..." link above.
200920102011201220132014201520162017201820192020202120222023
Google searches for 'no cap' (Rel. search volume)7.59.4166712.083317.666722.252024.2529.916734.916749.556.416783.91676754.7549.3
The average number of likes on Tom Scott's YouTube videos (Average likes)55743.321382627.6716317.363694.530265.739443.946468.269715.9103007101954209050169765132306107917




Why this works

  1. Data dredging: I have 25,153 variables in my database. I compare all these variables against each other to find ones that randomly match up. That's 632,673,409 correlation calculations! This is called “data dredging.” Instead of starting with a hypothesis and testing it, I instead abused the data to see what correlations shake out. It’s a dangerous way to go about analysis, because any sufficiently large dataset will yield strong correlations completely at random.
  2. Lack of causal connection: There is probably Because these pages are automatically generated, it's possible that the two variables you are viewing are in fact causually related. I take steps to prevent the obvious ones from showing on the site (I don't let data about the weather in one city correlate with the weather in a neighboring city, for example), but sometimes they still pop up. If they are related, cool! You found a loophole.
    no direct connection between these variables, despite what the AI says above. This is exacerbated by the fact that I used "Years" as the base variable. Lots of things happen in a year that are not related to each other! Most studies would use something like "one person" in stead of "one year" to be the "thing" studied.
  3. Observations not independent: For many variables, sequential years are not independent of each other. If a population of people is continuously doing something every day, there is no reason to think they would suddenly change how they are doing that thing on January 1. A simple Personally I don't find any p-value calculation to be 'simple,' but you know what I mean.
    p-value calculation does not take this into account, so mathematically it appears less probable than it really is.
  4. Confounding variable: 2020 is particularly different from the other years on this graph. Confounding variables (like global pandemics) will cause two variables to look connected when in fact a "sneaky third" variable is influencing both of them behind the scenes.




Try it yourself

You can calculate the values on this page on your own! Try running the Python code to see the calculation results. Step 1: Download and install Python on your computer.

Step 2: Open a plaintext editor like Notepad and paste the code below into it.

Step 3: Save the file as "calculate_correlation.py" in a place you will remember, like your desktop. Copy the file location to your clipboard. On Windows, you can right-click the file and click "Properties," and then copy what comes after "Location:" As an example, on my computer the location is "C:\Users\tyler\Desktop"

Step 4: Open a command line window. For example, by pressing start and typing "cmd" and them pressing enter.

Step 5: Install the required modules by typing "pip install numpy", then pressing enter, then typing "pip install scipy", then pressing enter.

Step 6: Navigate to the location where you saved the Python file by using the "cd" command. For example, I would type "cd C:\Users\tyler\Desktop" and push enter.

Step 7: Run the Python script by typing "python calculate_correlation.py"

If you run into any issues, I suggest asking ChatGPT to walk you through installing Python and running the code below on your system. Try this question:

"Walk me through installing Python on my computer to run a script that uses scipy and numpy. Go step-by-step and ask me to confirm before moving on. Start by asking me questions about my operating system so that you know how to proceed. Assume I want the simplest installation with the latest version of Python and that I do not currently have any of the necessary elements installed. Remember to only give me one step per response and confirm I have done it before proceeding."


# These modules make it easier to perform the calculation
import numpy as np
from scipy import stats

# We'll define a function that we can call to return the correlation calculations
def calculate_correlation(array1, array2):

    # Calculate Pearson correlation coefficient and p-value
    correlation, p_value = stats.pearsonr(array1, array2)

    # Calculate R-squared as the square of the correlation coefficient
    r_squared = correlation**2

    return correlation, r_squared, p_value

# These are the arrays for the variables shown on this page, but you can modify them to be any two sets of numbers
array_1 = np.array([7.5,9.41667,12.0833,17.6667,22.25,20,24.25,29.9167,34.9167,49.5,56.4167,83.9167,67,54.75,49.3,])
array_2 = np.array([55743.3,2138,2627.67,16317.3,63694.5,30265.7,39443.9,46468.2,69715.9,103007,101954,209050,169765,132306,107917,])
array_1_name = "Google searches for 'no cap'"
array_2_name = "The average number of likes on Tom Scott's YouTube videos"

# Perform the calculation
print(f"Calculating the correlation between {array_1_name} and {array_2_name}...")
correlation, r_squared, p_value = calculate_correlation(array_1, array_2)

# Print the results
print("Correlation Coefficient:", correlation)
print("R-squared:", r_squared)
print("P-value:", p_value)



Reuseable content

You may re-use the images on this page for any purpose, even commercial purposes, without asking for permission. The only requirement is that you attribute Tyler Vigen. Attribution can take many different forms. If you leave the "tylervigen.com" link in the image, that satisfies it just fine. If you remove it and move it to a footnote, that's fine too. You can also just write "Charts courtesy of Tyler Vigen" at the bottom of an article.

You do not need to attribute "the spurious correlations website," and you don't even need to link here if you don't want to. I don't gain anything from pageviews. There are no ads on this site, there is nothing for sale, and I am not for hire.

For the record, I am just one person. Tyler Vigen, he/him/his. I do have degrees, but they should not go after my name unless you want to annoy my wife. If that is your goal, then go ahead and cite me as "Tyler Vigen, A.A. A.A.S. B.A. J.D." Otherwise it is just "Tyler Vigen."

When spoken, my last name is pronounced "vegan," like I don't eat meat.

Full license details.
For more on re-use permissions, or to get a signed release form, see tylervigen.com/permission.

Download images for these variables:


View another random correlation

How fun was this correlation?

I'm grateful for your review!


Correlation ID: 4550 · Black Variable ID: 1368 · Red Variable ID: 25453
about · subscribe · emailme@tylervigen.com · twitter

CC BY 4.0