Home
Data Science

QS World University Rankings 2025 Analysis Using Python and Deepnote

In the era of globalization, university rankings have become an important aspect for prospective students, educational institutions, and policymakers. One of the most recognized international rankings is the QS World University Rankings. In this article, I will explain my analytical project on QS World University Rankings 2025 using Python and the interactive platform Deepnote, along with predictive analytics insights.

Live Demo | Source Code

QS World University Rankings 2025 Dataset Overview

The dataset analyzed was sourced from the QS World University Rankings 2025, containing rankings, scores from various indicators, and detailed information about universities globally. The key indicators used by QS for rankings are:

Research and Discovery (50%)

  • Academic Reputation (30%)
  • Citations per Faculty (20%)

Employability and Outcomes (20%)

  • Employer Reputation (15%)
  • Employment Outcomes (5%)

Global Engagement (15%)

  • International Faculty Ratio (5%)
  • International Students Ratio (5%)
  • International Research Network (5%)

Global Engagement (10%)

  • Faculty/Student Ratio (10%)

Sustainability (5%)

  • Sustainability (5%)

The dataset also includes categorical information such as:

  • SIZE: The size of the institution (e.g., XL, L, M, S)
  • FOCUS: The institutional focus (e.g., comprehensive (CO), focused (FO), specialist (SP), faculty-focused (FC))
  • RES.: Research intensity (e.g., very high (VH), high (HI), medium (MD), low (LO))
  • STATUS: Institutional status or category (e.g., A, B, C)

Data Preparation

The analysis began with data cleaning and preparation using Python libraries such as Pandas and NumPy. Key steps included:

  • Converting rank and score columns to numeric formats.
  • Handling missing data by filling gaps with median values or special categories.
  • Mapping categorical data (e.g., university size, institutional focus, research intensity) to numeric scales.

Descriptive Statistics and Data Exploration

I performed exploratory data analysis (EDA) using descriptive statistics to gain insights into data characteristics, such as average academic scores, employer reputation, and research intensity across various universities. This approach provided a clear understanding of the distribution of scores and helped identify significant trends within the dataset.

Interactive Visualization

A central part of the project involved interactive visualization created using Plotly on the Deepnote platform. The visualizations included:

  • Radar Charts comparing individual university performances against global averages.
  • Bar Charts providing detailed performance indicators.
  • Line Charts showing ranking trends over multiple years.
  • Parallel Categories visualizing score comparisons between universities across key indicators.
  • Choropleth Maps showing the global distribution of top-ranking universities.

These visualizations offered intuitive and interactive insights into university positions within a global context, allowing users to explore data dynamically.

Predictive Analytics

Additionally, predictive analytics was employed to forecast future rankings and performance trends based on historical data and indicator scores. Utilizing machine learning techniques, including regression models, predictions were made to anticipate potential shifts in university rankings, providing valuable foresight for strategic decision-making and university planning.

QS World University Ranking Performance Dashboard

As a final outcome, I developed an interactive dashboard enabling users to quickly access key information. This dashboard presents critical insights such as:

  • Top-ranked university (MIT with a perfect score of 100)
  • Specific university positions, e.g., Universitas Padjadjaran (ranked 517)
  • Countries with the highest number of top 100 universities (United States)
  • Average scores across essential indicators such as Academic Reputation, Citations per Faculty, and Sustainability

This project showcases not only the practical application of Python programming and data analysis but also demonstrates the powerful use of interactive visualization and predictive analytics in making data-driven decisions. I hope this analysis provides useful insights and encourages further exploration into leveraging modern analytical techniques for strategic academic planning.

If you are interested in purchasing the source code of this project, you can visit https://lynk.id/payme/sholehqomaruddin. The source code is fully modifiable and customizable. For custom analysis or further inquiries, feel free to contact me via the contact page.