A full AI-powered system built step-by-step through multiple milestones in Google Colab.
KnowMap analyzes text from different knowledge domains (Science → Business → Technology → etc.), extracts key concepts, generates semantic embeddings, and creates an interactive knowledge graph that shows cross-domain relationships.
This project was developed milestone-by-milestone and finally deployed using Streamlit + Ngrok.
KnowMap is an AI system designed and built by Alok Yadav.
It performs:
- NLP-based keyphrase extraction
- Sentence-Transformer embedding generation
- Knowledge graph creation using NetworkX + PyVis
- Cross-domain similarity linking
- Interactive visualisation through Streamlit
- Optional storage in Neo4j
This project was created entirely in Google Colab, with code executed milestone-by-milestone in clean modular cells.
- Upload multiple domain documents
- Preprocess text
- Remove stopwords, special characters
- Convert to clean tokenized text
- Use spaCy
"en_core_web_sm" - Extract Nouns, Noun-Phrases, Keywords
- Store domain-wise keyword lists
- Use Sentence-Transformer (
all-MiniLM-L6-v2) - Convert extracted concepts to embeddings
- Store embeddings for cross-domain comparison
- Build graph using NetworkX
- Add similarity-based edges
- Use PyVis for interactive graph visualisation
- Integrated all milestones into one app
- Launched via Ngrok from Colab
- User uploads multiple domain files, generates graph, and explores it live
Knowmap-cross-domain-knowledge-mapping-using-AI/
│
├── app.py # Final Streamlit App (All milestones integrated)
├── knowmap_cross_domain_knowledge_mapping_using_ai_.py # Colab exported code
├── requirements.txt # Libraries required
│
├── milestones/
│ ├── milestone_1_dataset.py
│ ├── milestone_2_nlp.py
│ ├── milestone_3_embedding.py
│ ├── milestone_4_graph.py
│ └── milestone_5_streamlit.py
---





