逢甲商圈周邊路邊停車壓力分析：以端午節連假期間為例

Abstract

中文摘要 本報告目的在於分析端午節連假期間逢甲商圈周邊路邊停車壓力的時間與空間變化，了解不同日期、時段及路段的占用情形，並檢查資料蒐集過程是否存在缺口或感測異常。研究使用臺中市政府路邊剩餘車位 API，每十分鐘蒐集一次車格狀態，先保存原始快照，再以 Azure VM、Docker Compose、collector 與TimescaleDB 建立持續蒐集及儲存環境。分析時以逢甲商圈中心（24.17950, 120.64630）為中心，篩選半徑一公里內的資料，依日期、時段及路段進行彙整。2026 年 6 月 18 日至 21 日共取得 337,536 筆有效觀測，涵蓋 25 個路段與 146 個車格。結果顯示，整體占用率為 69.8%，週六最高，達 71.5%。夜市時段最高，達 77.7%。路段 6458001 的高壓力時段佔比最高，達 88.2%。每小時序列的一小時落後自我相關為 0.90，顯示短期停車壓力具有延續性。預期抓取批次覆蓋率為100%，且未出現感測故障紀錄。雖然觀測期間只有四天，但仍可作為本次連假期間停車壓力描述、短期停車資訊規劃與後續分析的參考。

Abstract The purpose of this report is to analyze the temporal and spatial patterns of on-street parking pressure around Feng Chia Commercial District during the Dragon Boat Festival holiday, compare occupancy across dates, time periods, and road sections, and assess whether the data collection process contains gaps or sensor anomalies. Parking-space status was collected every ten minutes from the Taichung City Government open-data API. Raw snapshots were retained, and Azure VM, Docker Compose, a collector service, and TimescaleDB were used to establish the data collection and storage environment. Using the Feng Chia commercial district center (24.17950, 120.64630) as the reference point, we filtered parking spaces within a one-kilometer radius and aggregated the data by date, time period, and road section. From 18 to 21 June 2026, the study obtained 337,536 valid observations from 146 parking spaces across 25 road sections. The overall occupancy rate was 69.8%, with the highest daily rate observed on Saturday at 71.5%. The night-market period recorded the highest occupancy rate at 77.7%, and section 6458001 had the highest proportion of high-pressure periods at 88.2%. The one-hour lag autocorrelation was 0.90, indicating short-term persistence in parking pressure. All expected collection batches were completed, representing 100% batch coverage, and no sensor-failure records were observed. Although the observation period covered only four days, the findings can still serve as a reference for describing parking pressure during the holiday period, planning short-term parking information, and conducting further analysis.