AI explanation

^{Show image generation prompt} Model: dalle-3
Prompt: Generate an image of a solar farm in Brazil, with rows of panels stretching towards the horizon and absorbing sunlight. At the edge of the farm, a group of people are playfully dancing in a suspicious manner, casting elongated shadows across the ground. Overhead, a clear blue sky frames a hovering screen with the Google logo, and a search bar displaying the words "that is sus." 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: Solar power generated in Brazil and Google searches for 'that is sus'. 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: a meme Woodblock painting Surrealism a funny image macro Charcoal
ChatGPT: a funny image macro

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 Solar power generated in Brazil and Google searches for 'that is sus', in the following situation: The solar panels were accidentally programmed to emit a signal that sounds like the phrase 'that is sus,' leading to a viral TikTok trend of people dancing in a shady manner whenever they visited a solar farm.. Keep the description simple.
ChatGPT: Amidst the sprawling solar farm in Brazil, rows of panels stretch towards the horizon, absorbing the golden sunlight. At the edge of the farm, a group of people are seen dancing in a playful, suspicious manner, their movements casting elongated shadows across the ground. Meanwhile, overhead, the clear blue sky frames the occasional search bar, with the words "that is sus" and the Google logo on a hovering screen.

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 funny image macro ] and depicting this scene: Amidst the sprawling solar farm in Brazil, rows of panels stretch towards the horizon, absorbing the golden sunlight. At the edge of the farm, a group of people are seen dancing in a playful, suspicious manner, their movements casting elongated shadows across the ground. Meanwhile, overhead, the clear blue sky frames the occasional search bar, with the words "that is sus" and the Google logo on a hovering screen. Keep the visual depiction simple. Do not include any words or speech bubbles.
ChatGPT: Generate an image of a solar farm in Brazil, with rows of panels stretching towards the horizon and absorbing sunlight. At the edge of the farm, a group of people are playfully dancing in a suspicious manner, casting elongated shadows across the ground. Overhead, a clear blue sky frames a hovering screen with the Google logo, and a search bar displaying the words "that is sus."

---

*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.

AI academic paper

Spotlight on Solar Sus-picion: Illuminating the Relationship Between Solar Power Generation in Brazil and Google Searches for 'That is Sus' The Journal of Solar Sociolinguistics Jan 2024

Problem variable:	Which variable has a problem? Solar power generated in Brazil Google searches for 'that is sus' Neither / both
Issue:	What is the issue? The data is incorrect The chart isn't displaying properly There is a typo The source link is incorrect A calculation is wrong Something else
Additional details: Optional
Confirm you are a human:

Data details

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^Note 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^Note 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.
5. Outlandish outliers: There are "outliers" in this data.^Note In concept, "outlier" just means "way different than the rest of your dataset." When calculating a correlation like this, they are particularly impactful because a single outlier can substantially increase your correlation.
   
   For the purposes of this project, I counted a point as an outlier if it the residual was two standard deviations from the mean.
   
   (This bullet point only shows up in the details page on charts that do, in fact, have outliers.)
   They stand out on the scatterplot above: notice the dots that are far away from any other dots. I intentionally mishandeled outliers, which makes the correlation look extra strong.

Try it yourself

# 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([0.0001,0.0002,0.0002,0.0002,0.0002,0.0002,0.001,0.002,0.002,0.005,0.017,0.058652,0.084498,0.827103,3.44106,6.58803,10.6423,16.752,])
array_2 = np.array([3.83333,7.08333,8.66667,3.83333,1.33333,6.41667,5.58333,6.91667,10.25,10.0833,8.66667,8.66667,11.0833,12.5,15.4167,14.4167,40.8333,58.1667,])
array_1_name = "Solar power generated in Brazil"
array_2_name = "Google searches for 'that is sus'"

# 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

• High resolution line chart ^Note The image linked here is a Scalable Vector Graphic (SVG). It is the highest resolution that is possible to achieve. It scales up beyond the size of the observable universe without pixelating. You do not need to email me asking if I have a higher resolution image. I do not. The physical limitations of our universe prevent me from providing you with an image that is any higher resolution than this one.
  
  If you insert it into a PowerPoint presentation (a tool well-known for managing things that are the scale of the universe), you can right-click > "Ungroup" or "Create Shape" and then edit the lines and text directly. You can also change the colors this way.
  
  Alternatively you can use a tool like Inkscape.
  
• High resolution line chart, optimized for mobile
• Alternative high resolution line chart
• Scatterplot
• Portable line chart (png)
• Portable line chart (png), optimized for mobile
• Line chart for only Solar power generated in Brazil
• Line chart for only Google searches for 'that is sus'
• AI-generated correlation image
• The spurious research paper: Spotlight on Solar Sus-picion: Illuminating the Relationship Between Solar Power Generation in Brazil and Google Searches for 'That is Sus'