A wholesaler supplies clothing to thousands of stores and needs a model that predicts daily sales per item per store. Over half the stores are less than six months old. Sales are consistent week to week. The current dataset is weekly-aggregated and omits weeks with no sales. Five years of data (100 MB) are in cloud storage. Which issues are likely to harm forecast performance, and what actions should you take to address them? (Choose two.)

Detecting seasonality will be difficult for newer stores because they lack sufficient historical data. Requesting categorical attributes (e.g., store type, location) allows the model to leverage data from similar, older stores to infer seasonality patterns, improving predictions for new stores. Weekly aggregated data obscures daily fluctuations and patterns, which are crucial for accurate daily sales predictions. Obtaining daily sales records enables the creation of a more granular and accurate daily forecasting model. The other options are less impactful or incorrect. Obtaining external sales data from other industries is unlikely to improve performance for clothing sales. Requesting zero-entry records is important but less critical than daily data or addressing new store seasonality. Requesting more data (10 years) is not necessarily better if the data quality or granularity is poor; 100 MB of data over five years is substantial for this type of problem.