小黄姜线上渠道销售数据集

清洗异常值，按周、月等维度聚合时间序列数据，对销量、销售额进行标准化，计算各市州与区县平均销售曲线的相似度；利用K-means聚类算法将区县划分为高波动、稳定、低活跃3类，生成每个区域季节指数, 价格弹性系数, 增长斜率三元组；并绘制出区域销售特征图谱，实现时空可视化呈现，运用 ARIMA+LSTM 混合模型，对未来 30 天的销量进行预测。Clean outliers, aggregate time series data by dimensions such as week and month, standardize sales volume and sales revenue, and calculate the similarity between the average sales curves of each prefecture-level city and its subordinate districts and counties. Utilize the K-means clustering algorithm to classify districts and counties into three categories: high volatility, stable, and low activity, and generate the triplet consisting of seasonal index, price elasticity coefficient and growth slope for each region. Additionally, plot the regional sales feature map to achieve spatio-temporal visualization, and employ the hybrid ARIMA+LSTM model to forecast the sales volume for the next 30 days.

搜集汇总

背景与挑战

背景概述

小黄姜线上渠道销售数据集是一个2GB规模的数据集，每日更新，主要用于分析小黄姜在贵州各区域的销售情况，预测未来销售趋势，并支持差异化营销策略和政府农业补贴政策的制定。数据集包含时间、商品名称、市州、区县、销量和销售额等字段，通过动态时间规整（DTW）算法和ARIMA+LSTM混合模型进行数据分析和预测。