De Heus Cassava price forecasting

During the Capstone project, the Big O(no) team implemented the website called Cassava Price Prediction. The website is used to predict cassava prices across 15 existing factories using machine learning models such as XGBoost, NLinear, and DLinear. The website includes two line graphs to visualize the cassava price predictions and historical gas prices. By analyzing historical cassava and gas prices, users can generate cassava price forecasts for the next six timestamps. The website also provides in-depth explanations of these predictions through llama, mixtral, and gemma LLM modules for further understanding. Moreover, users can upload data files in XLS, or XLSX formats, which will be stored locally, with a preview option available before predictions. The data file must follow a specific format, including the columns: time, price (representing historical cassava prices), and gas_price. The website also supports theme customization, including light, dark, and system themes. User authentication is required, with secure access via username and password. For admin, admin can manage accounts, including creating, editing, deleting accounts, as well as searching for existing accounts. In conclusion, this website is tailored for De Heus company, providing precise and interpretable cassava price predictions with an easy-to-use interface.