Forecasting Daily Number of User Problem Reports of Junyi Academy for Efficient Staff Allocation

Project Details


Fall 2016


Chia Li Chien, Sherry Wu, Elisa Wang, Emily Wu





Introduction of Our Client: Junyi Academy In this project, our client is Junyi Academy, a platform offering online learning resources for all ages. It provides practice exercises, instructional videos, and a personalized learning dashboard that empower learners to study at their own pace in and outside of the classroom. With high utilization ratio of the practice exercises, Junyi receives problems reported by users, which are called “user problem reports”. All the reports will be checked and then be distributed to the responsible team by operation team.

Business Problem: The main business goal for this project is to help the manager of Junyi Academy operation team better allocate the staffs and their work loading. Since there is no full-time staff dealing with user problem reports, if daily reports are over 23, which is the average number of reports solved, it is very likely that reports cannot be solved on that day. Our project is going to forecast daily number of user problem reports of the next week. With the forecasts, the manager of operation team can come up with proper actions to handle the days with plenty reports.

Data Description: We got data from Junyi Academy, including daily number of user problem reports, daily number of active users and daily number of new registered users. The time period of the series are from Aug. 29th, 2016 to Nov. 13th, 2016, totally 77 records in each series. We also marked out the “student school day” as a series, which meant whether students had to go to school or not, and also “outlier”, which meant the usual number of the user problem reports.

Forecasting Solution: Before forecasting, we applied data visualization technique to detect data patterns. Then, we chose and try several forecasting methods appropriate for our data, including seasonal naive, moving average, Holt-Winter’s smoothing (ANA), linear regression model and neural networks. Finally, considering both our goal and prediction accuracy, we decided to apply neural networks and also do roll-forward forecasting to get the prediction interval.

Recommendations: According to the results, we suggested Junyi Academy could just run our predictive model on every Friday to forecast daily number of user problem reports of the next week. Thus, the manager of operation team can, based on the forecasts and prediction interval, decide whether to allocate extra staff to deal with reports or to arrange people to check the content of the questions before released which is the main cause for user problem reports.

Application Area: