Forecasting swimsuit sales for the next month to assist inventory management for Heatwave Team 4 : Jheng Kai-Ru (107078503) Adam Yu (107078506) Silvia Yang (107078507) Zoly Chang (107078509)
Business Problem Client Problem Sell on Heatwave has limited knowledge toward decision on how many product to produce; also to put promotions on certain products. They mostly do it base on their experiences. Sometimes it works, Heatwave is a B2C swimsuits seller on a sometimes it doesn’t. Chinese e-commerce platform called Goal TMall. They design and manufacture their own swimsuits Using the historical sales data to forecast the sales of the next month to assist with their inventory management and production strategy.
Data Description: Data Constraints Problems ● Data from Shen-Yi-Can-Mou is too short! ● Order doesn’t match in order list and item list ● Some products may have different product name, and may have different product id ● Two sources of data have different product id ● We miss 1314 data when integrated them Although the data are not very accurate, we think it will still be helpful for our forecasting : )
Data Description: After Preprocessing Daily Data Monthly Data From 2017-05 to 2018-11 From 2017-04-22 Summer has higher sales to 2018-12-18 Holiday has higher sales
Data Description: Original Data Shen-Yi-Can-Mou Compute date Product ID Product Name # of payments $ of payments ● Data : Monthly and daily Sales data ● Time Period : Daily 2018/07/25 - 2018/12/18 Monthly 2017/09 - 2018/11 ● Data Quality : good Item List Order List TMall’s Analytic Platform Order Product ID Order # of products ● Data : Daily data, contain order list and 1 82275 item list 1 3 1 84567 ● Time Period : 2017/04/22 - 2018/11/24 2 2 ● Data Quality : bad 1 87632 3 1 2 82275 2 87632
Data Preprocessing TMall Analytics Order List Item List Find the product names and Separate product product id, if there has same name names into one row but different id, change the id. Join two list Find the pId No If there from pName has pId Yes No We cannot know If there which product it is has pId Aggregate sales data into daily We have 1314 row data don’t Sheng-Yi-Can-Mou Data and monthly data frame Yes have pId, so we ignore it. Merge into the data frame Forecasting data
Method: Forecast Monthly Data P82275 monthly forecast error (RMSE) Model Training error Test error sNaive 336.8237 93.0430 regression 129.15176 76.34134 arima 289.2212 222.6107 Ensemble 251.7322 130.665
Method: Forecast Daily Data P82275 daily forecast error (RMSE) Training error Test error sNaive 11.582822 2.632218 ets 8.317506 2.138090 arima 8.317509 2.138103 Ensemble 9.405945667 2.302803667
Evaluation Monthly (RMSE) Daily (RMSE) 121.4645 4.53995822 sNaive (Benchmark) 38.87665 1.50777218 131.9892 3.315389 Overfitting! ets 174.2751 1.00362794 40.12895 5.535291 regression 41.63898 3.6440104 102.0755 3.294651 arima 99.33205 0.98115324 87.88965 3.716666073 Ensemble Over-forecast! (top 3 model) 59.9492265 1.164184453
Recommendations & Limitations Recommendation 1. sNaive for monthly, ets regression for daily. 2. Retain their data autonomy , and confirm the data quality additionally. 3. C ompare the similarities. 4. Data long enough inventory management reaching lean production . 5. Forecasting + Domain Knowledge Constraint Monthly 1. Forecasting future clothing trend is hard when solely using the data. 2. Short product life cycle. 3. Data constraint
Recommend
More recommend
