This project explores the CORD-19 dataset (metadata of COVID-19 research papers).
We perform:
- Data cleaning & preparation
- Exploratory data analysis
- Visualizations (publication trends, top journals, word cloud of titles)
- An interactive Streamlit web application for exploration
The analysis is based on a sample dataset (sample_metadata.csv) created for testing.
pip install -r requirements.txt
pip install pandas matplotlib seaborn streamlit wordcloud
📊 Features 🔹 Data Cleaning
Handles missing values in critical columns (title, publish_time)
Converts publish_time to datetime format
Extracts year column for analysis
Adds abstract_word_count column
🔹 Visualizations
Bar chart: Publications by year
Horizontal bar chart: Top 10 journals
Word cloud: Common words in paper titles
🔹 Streamlit App
Interactive slider to filter year range
Dynamic visualizations (publications over time, top journals)
Displays sample data in a table
📷 Example Outputs
Publications trend chart
Top journals chart
Word cloud visualization
Streamlit app interface with filters
📝 Reflection
Challenges faced:
Handling missing and inconsistent data
Managing dataset size (using sample dataset for testing)
Installing extra libraries (wordcloud, streamlit)
What I learned:
End-to-end workflow of data science (cleaning → analysis → visualization → app)
Basics of Streamlit for building interactive dashboards
Using WordCloud for quick text-based insights
📌 Requirements
Python 3.7+
pandas
matplotlib
seaborn
wordcloud
streamlit