Inspiration : Need to analyze and understand complex user interactions in large-scale networks, such as the Wikipedia Talk Network.
What it does : Processes the Wikipedia Talk Network to identify influential users, detect communities, and understand communication patterns.
How we built it : Downloaded the Wikipedia Talk Network dataset from SNAP.Loaded the dataset into a NetworkX graph and persisted it in ArangoDB.Used LangChain and LangGraph to build the agent.Integrated OpenAI's GPT models for natural language query processing.Created custom tools like graph_qa_tool and cugraph_analysis_tool to handle graph queries and analytics.Used NetworkX and Matplotlib to visualize a sample of the graph.Built a Gradio interface for users to interact with the application.
Challenges we ran into:The dataset contains 2.3 million nodes and 5 million edges, making it challenging to load and process efficiently.Solved by loading the graph in batches and using GPU-accelerated analytics with cuGraph.
Accomplishments that we're proud of : Scalable Graph Analytics,User-Friendly Interface.
What we learned:Learned how to analyze large-scale graphs using tools like NetworkX, cuGraph, and ArangoDB
What's next for Next-Gen Agentic App : Add support for other graph datasets, such as social networks or citation networks.Add more tools for advanced graph analytics, such as community detection and shortest path algorithms.
