共享单车供给决策预警数据

进行共享单车单位管理区划分，建立单位管理区域泊位占有率预测模型，解决共享单车企业对用户需求预测不准确、资源调配不合理导致市场供需失衡的问题，从而实现向共享单车运营平台发送供给量决策预警，缓解用户借车难、还车难的问题。1、数据采集处理：获取管理区域内共享单车用户需求行为的特征数据以及每个单位管理区域内t-1时刻的共享单车使用数据。根据特征数据将管理区域划分为多个共享单车单元管理区，预设单元管理区D划分规则为：距离轨道站点半径范围：D1＜600m，D2：600~1000m，D3＞1000m。2、数据分析：根据距离轨道站点的可达性和区域用地类型确定各单元管理区的泊位占有率决策阈值，决策阈值分为高限阈值θh和低限阈值θl两种。第一类单元管理区D1，过高预警阈值1，过低预警阈值0.4；第二类单元管理区D2，过高预警阈值0.9，过低预警阈值0.5；第三类单元管理区D3，过高预警阈值0.8，过低预警阈值0.2。根据预测模型yt＝ax+m，a、m为模型参数，x为t-1时刻的泊位占有数，yt为t时刻泊位占有数。预测泊位占有率θ=预测泊位占有数yt/泊位总数，判断各单元管理区的共享单车的预测泊位占有率是否超过设定的决策阈值，若是，则向调度平台发出预警信号。

This work conducts the division of bicycle-sharing unit management zones and constructs a parking spot occupancy rate prediction model for unit management zones, aiming to address the market supply-demand imbalance caused by inaccurate user demand forecasting and unreasonable resource allocation of bicycle-sharing enterprises, so as to send supply decision early warnings to bicycle-sharing operation platforms and alleviate the difficulties of users in borrowing and returning bicycles. 1. Data Collection and Processing: Collect the characteristic data reflecting the demand behavior of bicycle-sharing users within the management zones, as well as the bicycle-sharing usage data of each unit management zone at time t-1. Divide the overall management zones into multiple bicycle-sharing unit management zones based on the collected characteristic data. The preset division rules for unit management zones D are specified as: radius range from rail transit stations: D1 < 600m, D2: 600–1000m, D3 > 1000m. 2. Data Analysis: Determine the decision thresholds for parking spot occupancy rate of each unit management zone based on its accessibility to rail transit stations and regional land use types. The decision thresholds are divided into two types: upper limit threshold θ_h and lower limit threshold θ_l. For the first type of unit management zone D1, the upper early warning threshold is 1 and the lower early warning threshold is 0.4; for the second type D2, the upper threshold is 0.9 and the lower threshold is 0.5; for the third type D3, the upper threshold is 0.8 and the lower threshold is 0.2. Using the prediction model $y_t = a x + m$, where $a$ and $m$ are model parameters, $x$ is the number of occupied parking spots at time t-1, and $y_t$ is the number of occupied parking spots at time t. Calculate the predicted parking spot occupancy rate as $ heta = y_t / ext{total number of parking spots}$, then judge whether the predicted parking spot occupancy rate of bicycle-sharing in each unit management zone exceeds the set decision thresholds. If so, send an early warning signal to the scheduling platform.

搜集汇总

数据集介绍

特点

该数据集由宁波喵走科技有限公司提供，包含2001条数据，每月更新一次。数据集用于共享单车单位管理区划分和泊位占有率预测，旨在解决共享单车供需失衡问题，提升资源调配效率。