Post by Kailin (Brian) Chang
Project Manager at MongoDB
做了一個自己開心的小玩具 Demo,MLB 棒球數據分析,以 MongoDB Atlas 為核心開發,目前抓四個球季資料與部分新聞,每日更新。 原本三月球季開打前就想做的,拖到都球季中,連世足都開踢了,一度想說改做世足 😂。 🔗 做開心的,只是 Demo: https://lnkd.in/gc8G-eke 功能: 👤 球員數據 — 復古球員卡、AI 球探報告、百分位排名、歷年走勢圖 🧬 相似球員 — DNA 雷達圖視覺化球員特徵,Atlas Vector Search 向量搜尋相似球員 📰 MLB 新聞 — Atlas Automated Embedding + Voyage AI 新聞語意搜尋 💬 AI 查詢 — Agent Tools / Text-to-Query 雙模式,自然語言直接問資料庫 📊 排行榜、分區戰績、球員比較、球隊對比⋯⋯等共九個頁面 MongoDB Atlas 讓這一切變簡單: • Flexible Schema — 投手、打者統計欄位天生不同,資料模型彈性擴充 • Atlas Vector Search — 球員統計數據向量化,找出打法相似的球員 • Atlas Search — 球員姓名模糊搜尋、自動補全,支援拼字容錯 • Atlas Automated Embedding + Voyage AI — 新聞寫入即自動建立向量,無需額外 embedding pipeline 功能整合在同一個平台,不需要額外的向量資料庫或搜尋服務。 ⚠️ 服務不定期關站,目前都用免費方案,AI 問太多Token 就爆掉啦 😂 Tech Stack: MongoDB Atlas · Voyage AI · Gemini · Groq · Streamlit · Python ========================================================= Built a small MLB analytics demo, powered by MongoDB Atlas — 4 seasons of data and some news, updated daily. 🔗 Demo Just For Fun: https://lnkd.in/gc8G-eke Features: 👤 Player Dashboard — vintage player card, AI scouting report, percentile rankings, career trends 🧬 Similar Players — DNA radar chart + Atlas Vector Search to find players with the same playstyle 📰 MLB News — Atlas Automated Embedding + Voyage AI semantic news search 💬 AI Query — Agent Tools / Text-to-Query dual mode, ask the database in natural language 📊 Leaderboards, standings, player comparison, team matchups... 9 pages total MongoDB Atlas made this easy: • Flexible Schema — pitchers and hitters have different stat fields, no predefined schemas needed • Atlas Vector Search — vectorized player stats to find similar players • Atlas Search — fuzzy name search with autocomplete and typo tolerance • Atlas Automated Embedding + Voyage AI — auto-embed on write, no extra embedding pipeline One platform. No separate vector DB. No extra search service. ⚠️ Everything runs on free tier. Uptime not guaranteed, tokens not unlimited 😂