📊 Scientific Research Trends Analysis

🧠 Problem Statement

The growth of scientific publications mirrors the pace of global innovation, collaboration, and knowledge dissemination. Over the past few decades, open-access repositories like arXiv and platforms like Semantic Scholar have captured millions of research articles across disciplines.

This project explores how research fields expand, how collaboration networks evolve, and how citation patterns reflect scientific influence. By analyzing large-scale research datasets, we aim to uncover how science is being created, shared, and valued.

---

📂 Dataset

• Semantic Scholar API: https://www.semanticscholar.org/product/api
• arXiv Dataset (Cornell University): https://www.kaggle.com/datasets/Cornell-University/arxiv

---

📘 Project Overview

All code and visualizations are implemented in the Jupyter notebook:
Research_analytics.ipynb

---

🧾 Exploratory Questions

Below are the key exploratory research questions addressed in this project:

---

✅ Completed Analysis

1️⃣ Research Growth Trends

Question:
Which research areas (e.g., AI, medicine, renewable energy, climate science) show the fastest growth in publications, and are there bursts of activity following major breakthroughs?

Findings:

• AI and Medicine exhibit the fastest growth in publications since 2015.
• Noticeable bursts of activity:
  • AI: 2017–2018 (following deep learning breakthroughs).
  • Medicine: 2020 (COVID-19 related research).
• Renewable energy and climate science have demonstrated steady linear growth since 2010.

Visualization:
Publication trends across major disciplines.

---

3️⃣ Interdisciplinary Citation Advantage

Question:
Are interdisciplinary papers (tagged with multiple fields) cited more frequently than single-discipline papers?

Findings:

• Interdisciplinary papers receive ~1.4× more citations on average than single-discipline papers.
• Strongest cross-domain intersections:
  • AI + Medicine → medical imaging, drug discovery.
  • AI + Climate Science → predictive modeling, sustainability.
• This shows that cross-domain collaboration enhances visibility and citation impact.

Visualization:
Average citation comparison between interdisciplinary and single-discipline papers.

---

🕓 Work in Progress

• 🌍 Question 2: International collaboration network analysis.
• ⏳ Question 4: Citation half-life evolution.
• 🔑 Question 5: Emerging topic detection using TF-IDF and topic modeling.

---

🧰 Tech Stack

• Language: Python
• Notebook: Jupyter (Research_analytics.ipynb)
• Libraries:
  pandas, numpy, matplotlib, seaborn, plotly, nltk, scikit-learn

---

📈 Future Goals

• Build interactive dashboards using Plotly Dash or Streamlit.
• Integrate network analysis graphs for author and keyword co-occurrence.
• Publish a research report or paper based on findings.