United Nation's Goal of Achieving Zero Hunger By 2030

Overview

This project analyzes global Twitter conversations around the UN’s Sustainable Development Goal 2 (Zero Hunger) using Natural Language Processing (NLP). With data from 2019–2024, it leverages techniques like n-gram analysis, sentiment tracking, topic modeling, and emotion detection to uncover what the world is saying and feeling about food security, sustainability, and hunger. This project was made a student team, led by me, named "rookie.ai"

Tools and Technologies

Languages/Frameworks: Python, Hugging Face Transformers

Libraries: pandas, matplotlib, nltk, gensim, scikit-learn

Techniques:

• N-Grams (Unigrams, Bigrams, Trigrams)

• Time Series Tweet Count Analysis

• Sentiment Over Time

• LDA Topic Modeling

• Emotion Detection using Hugging Face (T5 fine-tuned model)

Part 1: N-Gram Analysis - What the World Talks About 

By analyzing unigrams, bigrams, and trigrams, we identified recurring keywords and themes.

Highlights:

• Frequent mentions: "food security", "malnutrition", "sustainable agriculture"

• These phrases reflect urgent global focus areas: not just hunger, but how climate change and agriculture impact it.

 Strategic Insight:

• These findings reveal a global framing of hunger as a systemic issue, with intertwined environmental and economic implications.

Part 2: Tweet Count Over Time – When People Speak Up 

We tracked tweet volume from 2019 to 2024, with drill-down into daily activity for July 2024, aligning with the UN High-Level Political Forum.

Key Findings:

• Sharp spike in mid-July 2024, clearly showing that major international events spark global engagement on Twitter.

• Activity around food security discussions mirrors the event calendar of global summits, policy launches, and public crises.
  
  

Strategic Insight:

• Governments and NGOs can time campaigns, awareness drives, and announcements to align with global forums for maximum reach.

Part 3: Sentiment Over Time – How People Feel About Hunger 

We tracked tweet volume from 2019 to 2024, with drill-down into daily activity for July 2024, aligning with the UN High-Level Political Forum.

Key Findings:

• Sharp spike in mid-July 2024, clearly showing that major international events spark global engagement on Twitter.

• Activity around food security discussions mirrors the event calendar of global summits, policy launches, and public crises.
  
  

Strategic Insight:

• Governments and NGOs can time campaigns, awareness drives, and announcements to align with global forums for maximum reach.

Part 4: LDA Topic Modeling – Discovering What’s Trending 

We used Latent Dirichlet Allocation (LDA) to find underlying discussion themes across the dataset.

Discovered Topics:

• Sustainable Agriculture & Food Security

• Malnutrition & Global Hunger

• Climate Impact on Food Systems

• Policy Change & Global Alliances

 Strategic Insight:

• Hunger is not discussed in isolation — it’s intertwined with climate change, economic stability, and global cooperation.

• Highlights a need for integrated policy solutions.

Part 5: Emotion Detection – What Emotions Drive These Conversations  

Using Hugging Face’s T5-base fine-tuned emotion model, we categorized tweets into:
Joy, Sadness, Anger, Fear, Love, Surprise

Findings:

• Joy dominated — often tied to campaigns, celebrations, or project launches

• Sadness & Fear appeared around food scarcity, famine, or policy failure

• Anger indicated public dissatisfaction with the political or social handling of hunger

Strategic Insight:

• Emotional resonance can be leveraged in communication campaigns.
  Joyful messaging and success stories amplify engagement, while fear/sadness highlights urgent needs for action and empathy.