<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Web | Daeun's Portfolio</title><link>https://ddaaan.github.io/en/tag/web/</link><atom:link href="https://ddaaan.github.io/en/tag/web/index.xml" rel="self" type="application/rss+xml"/><description>Web</description><generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Wed, 01 Jan 2025 00:00:00 +0000</lastBuildDate><image><url>https://ddaaan.github.io/media/logo_hu25717422b3b0790f0566abe2caaf12f9_60927_300x300_fit_lanczos_3.png</url><title>Web</title><link>https://ddaaan.github.io/en/tag/web/</link></image><item><title>Integrated Data Analysis and Automated Report Generation System</title><link>https://ddaaan.github.io/en/project/capston/</link><pubDate>Wed, 01 Jan 2025 00:00:00 +0000</pubDate><guid>https://ddaaan.github.io/en/project/capston/</guid><description>&lt;h2 id="-project-overview">💡 Project Overview&lt;/h2>
&lt;p>This system provides an &lt;strong>end-to-end data analysis service&lt;/strong> that allows users to upload their data and perform &lt;strong>search, automatic preprocessing, preview, RAG-based chatbot Q&amp;amp;A, and automatic Korean report generation&lt;/strong>.&lt;br>
By integrating Streamlit and Flask, it delivers both a user-friendly interface and powerful backend analytical capabilities.&lt;/p>
&lt;hr>
&lt;h2 id="-key-features">🌟 Key Features&lt;/h2>
&lt;ul>
&lt;li>&lt;strong>Data Search&lt;/strong>: Enables fast data retrieval using an integrated database.&lt;/li>
&lt;li>&lt;strong>Upload and Analysis&lt;/strong>: Users can upload and analyze their own datasets.&lt;/li>
&lt;li>&lt;strong>Automatic Preprocessing &amp;amp; Preview&lt;/strong>: Handles missing values and data type preprocessing automatically, with schema and sample data preview.&lt;/li>
&lt;li>&lt;strong>RAG-based Chatbot&lt;/strong>: Supports natural language Q&amp;amp;A using the user-selected file as the context.&lt;/li>
&lt;li>&lt;strong>Automatic Korean Report Generation&lt;/strong>: Summarizes LLM responses into a structured Korean report available for download.&lt;/li>
&lt;/ul>
&lt;hr>
&lt;h2 id="-tech-stack">⚙️ Tech Stack&lt;/h2>
&lt;ul>
&lt;li>&lt;strong>Language&lt;/strong>: Python&lt;/li>
&lt;li>&lt;strong>Framework&lt;/strong>: Streamlit, Flask&lt;/li>
&lt;li>&lt;strong>Server/Infra&lt;/strong>: Docker&lt;/li>
&lt;li>&lt;strong>Models&lt;/strong>: KURE-v1 (Embedding), Qwen3 (LLM)&lt;/li>
&lt;/ul>
&lt;hr>
&lt;h2 id="-project-highlights">🚀 Project Highlights&lt;/h2>
&lt;ul>
&lt;li>&lt;strong>RAG Architecture&lt;/strong>: Uploaded files are stored in S3, parsed, preprocessed, and embedded (via KURE-v1) for vector-based retrieval.&lt;/li>
&lt;li>&lt;strong>LLM Integration&lt;/strong>: Context generated from vector search is passed to the Qwen3 model for response summarization and rendered into a report template.&lt;/li>
&lt;li>&lt;strong>Docker-based Deployment&lt;/strong>: The project is containerized for seamless deployment across local and server environments.&lt;/li>
&lt;li>&lt;strong>Comprehensive Troubleshooting Documentation&lt;/strong>: Detailed notes on Docker setup and issue resolution were recorded.
&lt;blockquote>
&lt;p>🔧 &lt;a href="https://www.notion.so/Docker-Trouble-Shooting-1d284dccb3788007bf86c6856df9e8e6" target="_blank" rel="noopener">Docker Setup / Troubleshooting Guide&lt;/a>&lt;/p>
&lt;/blockquote>
&lt;/li>
&lt;/ul>
&lt;hr>
&lt;h2 id="-github-repository">🔗 GitHub Repository&lt;/h2>
&lt;p>&lt;a href="https://github.com/Ddaaan/capston" target="_blank" rel="noopener">https://github.com/Ddaaan/capston&lt;/a>&lt;/p></description></item></channel></rss>