Along with the increasing public demand for car transportation modes, car sales businesses are also increasing. Efforts to exist and be competitive are carried out such as by applying machine learning models to determine the car’s selling price based on its specification. Businesses can also stimulate sales by actively offering customers. The effectiveness of the active and massive offerings can be increased by personalizing the offers provided. This research uses a machine learning-based approach to learn customer profile data to predict the car’s price they would buy. The research was conducted by adopting the CRISP-DM framework and developed using the Google Colaboratory and Azure Machine Learning platforms. The modeling stage developed six regression models, those are linear regression, Lasso, Ridge, Random Forest Regressor, Elastic-net, and Support Vector Regressor (SVR). After the evaluation stage, the Lasso regression model with the performance of R-squared (R2) of 0,99958 and Mean Absolute Error (MAE) of 2.284.865,29 deployed as a web service endpoint so it could be accessed in real-time. The web service required the customer’s “Gender, Age, Annual Salary, Credit Card Debt, and Net Worth” to return a response of the recommended car price range prediction for the customer to buy. In further development, predictions obtained through web services can be implemented in public applications to display personalized car sales offers or pages based on customer profiles

Data is an important asset and a fundamental requirement for building valuable information for organizations. Association of Information Systems Students of Unsika (Himsika) as a university organization provides many events to develop student’s academic and professional skills. A post-event evaluation through a feedback survey was conducted and stored in Google Sheets spreadsheet format. However, the current analysis process using spreadsheets lacks standardization, making it difficult to compare satisfaction rates over time and between events. Additionally, the lack of standardization leads to semi-structured data on spreadsheets, with varying question formats and meanings. To address these limitations, implementing a centralized data warehouse is proposed as a solution. The data warehouse would provide a structured and standardized approach to analyzing event feedback, enabling better comparisons and evaluation of management quality within Himsika. The research aims to design a data warehouse that supports multidimensional analysis. As a way to simplify and optimize analytical queries, the data structure is standardized in the data warehouse. The Four-step Dimensional Design method is applied in designing dimensional modeling on the data warehouse, consisting of four stages including selecting the business process, declaring the grain, identifying the dimensions, and identifying the facts. The design process resulted in 4 dimensions of events, dim_instances, dim_degree_programs, and dim_professions, and a fact table called fact_rates_by_responses. Overall, the proposed data warehouse and dimensional modeling approach aim to enhance the analysis and evaluation of Himsika’s events.

The development of distribution and market segmentation has become the company's background in improving business processes. The purpose of this research is to analyze the business processes of beverage companies using Business Process Management (BPM) modeling and improvised based on six core element management. In the analysis process, it is found that there is no stock forecasting system in forecasting sales stock that must be fulfilled. The results of the study show that the Business Process Management model is improved with the addition of a stock forecasting system, so that business processes become more controlled with the presence of a product stock inventory forecasting system in the company.