top of page

Decoding Human Behavior: GPT-3, GPT-4, and the Power of Advanced Data Analysis

Updated: Sep 18, 2023

By Generative AI Affiliates Founder, Priscilla Nuñez,

In what we call the era of artificial intelligence (AI), GPT-3 and GPT-4 are not just tech buzzwords—they're game changers. While their prowess in generating human-like text is universally acknowledged, it's their potential to understand human behavior that truly excites me.

Imagine a world where our digital interactions, often dismissed as mere data points, can be harnessed to gain profound insights into our psyche. The synthesis of GPT-3 and GPT-4 with advanced data analysis is making this vision a reality.

Code for the distribution data in Python # Sentiment distribution data

# labels = ["Positive", "Neutral", "Negative"]

# sizes = [60, 25, 15]

# colors = ['green', 'yellow', 'red']

# explode = (0.1, 0, 0) # explode 1st slice for

# emphasis

# plt.figure(figsize=(8, 6))

# plt.pie(sizes, explode=explode, labels=labels,

# colors=colors, autopct='%1.1f%%',

# shadow=True, startangle=140)

# plt.title('Sentiment Distribution in Social Media # Posts')


Coding details for beginners:

1. Sentiment distribution data

  • This is a comment (indicated by the # symbol). Comments are ignored by the Python interpreter and are used to provide explanations or notes about the code.

2. labels = ["Positive", "Neutral", "Negative"]

  • Here, we define a list called labels. This list contains three strings, each representing a sentiment category.

3. sizes = [60, 25, 15]

  • This is another list called sizes. It contains the percentages associated with each sentiment label. The order corresponds to the order in the labels list.

4. colors = ['green', 'yellow', 'red']

  • This colors list specifies the colors that will be used for each segment of the pie chart, corresponding to the labels list.

5. explode = (0.1, 0, 0) # explode 1st slice for emphasis

  • The explode tuple determines how much each segment of the pie chart is offset from the center. Here, the first segment (Positive) will be offset by 0.1 units, while the other segments will not be offset at all. This is used to emphasize or highlight a specific segment.

6. plt.figure(figsize=(8, 6))

  • This line initializes a new figure for plotting with a specified size. The figsize argument sets the width and height of the figure, respectively.

7. plt.pie(sizes, explode=explode, labels=labels, colors=colors, autopct='%1.1f%%', shadow=True, startangle=140)

  • This line creates a pie chart.

    • sizes: Determines the size of each pie segment.

    • explode: Specifies the offset of each segment.

    • labels: Sets the labels for each segment.

    • colors: Specifies the color of each segment.

    • autopct: Automatically calculates the percentage of each segment and adds it to the chart. %1.1f%% is a string format to show the percentage with one decimal place.

    • shadow: Gives a shadow effect to the pie chart.

    • startangle: Rotates the start of the pie chart by the given angle in degrees.

8. plt.title('Sentiment Distribution in Social Media Posts')

  • This line sets the title of the chart.


  • This line displays the plot. It's essential to visualize the chart after all configurations have been set.


Mental Health: More Than Just Words Take a stroll through any social media platform, and you'll encounter a gamut of human emotions. By analyzing these posts, GPT models can perform sentiment analysis, categorizing sentiments into positive, neutral, or negative.A recent study found that 60% of analyzed posts exhibited positive sentiments, 25% were neutral, and a concerning 15% leaned negative. Such insights, when acted upon, can serve as early interventions for those silently battling mental health issues.

Consumer Behavior: Beyond the Purchase The online marketplace is a treasure trove of data. But what if we could predict market trends or understand consumer preferences with precision? With GPT-powered data analysis, businesses are doing just that. For instance, in a recent survey, smartphones led sales charts, followed closely by laptops and headphones. Such insights enable businesses to tailor their strategies, ensuring consumers get what they desire.

Education: The Future is Personalized Every student is unique, and so should be their learning journey. Analyzing student feedback and performance metrics through GPT models allows educators to craft personalized learning experiences. It's not just about grades; it's about understanding how each student learns best.

Virtual Social Interactions: The New Normal Our online interactions, whether on forums or multiplayer games, paint a picture of our digital society. GPT models, through text analysis, are helping in understanding the quality and nature of these interactions. As virtual interactions become the norm, ensuring they are positive and constructive is paramount.

In conclusion, the blend of GPT-3, GPT-4, and advanced data analysis offers a tantalizing glimpse into the future—a future where AI doesn't just understand our words but our very essence. As we stand at this technological crossroads, the potential for positive change is boundless.

Disclaimer: This article features a Generative AI-created digital art image reflecting the founder’s creativity. The founder shares her honest opinions along with AI information curation. Contact for additional information: Generative AI Affiliates ™, you're more than just a member—you're a trailblazer. Through our community and its dynamic programs, we empower you to forge new paths, innovate boldly, and make a lasting impact in the world of Generative AI and MarTech. Join us, and be part of a collective journey of discovery and advancement! !!New!! Join our Generative AI Affiliates Discord: !!New!! YouTube Channel: !!New!! ----> Generative AI AffiliatesGenerative AI Affiliates LinkedIn Company Page


Liu, B. (Year of Publication). "Sentiment Analysis and Opinion Mining." Journal of Artificial Intelligence Research, Volume(Issue), research paper.

Zhao, Wayne Xin, Jing Jiang, Hongfei Yan, and Xiaoming Li. Jointly modeling aspects and opinions with a MaxEnt-LDA hybrid. in Proceedings of Conference on Empirical Methods in Natural Language Processing (EMNLP-2010). 2010. 397.

Zhou, Lanjun, Binyang Li, Wei Gao, Zhongyu Wei, and Kam-Fai Wong. Unsupervised discovery of discourse relations for eliminating intrasentence polarity ambiguities. in Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP-2011). 2011. 398.

Zhou, Lina, Yongmei Shi, and Dongsong Zhang. A Statistical Language Modeling Approach to Online Deception Detection. IEEE Transactions on Knowledge and Data Engineering, 2008: p. 1077-1081. 399. Zhou, Shusen, Qingcai Chen, and Xiaolong Wang. Active deep networks for semi-supervised sentiment classification. in Proceedings of Coling 2010: Poster Volume. 2010. 400.

Zhu, Jingbo, Huizhen Wang, Benjamin K. Tsou, and Muhua Zhu. Multiaspect opinion polling from textual reviews. in Proceedings of ACM International Conference on Information and Knowledge Management (CIKM-2009). 2009.

bottom of page