Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.Accurate sales forecasting is essential for effective inventory management, pricing strategies, and operational planning in the retail sector. However, many small and medium-sized enterprises lack access to affordable and intelligent forecasting tools. This thesis presents the development of a mobile-based Android application for sales forecasting, integrating statistical and machine learning techniques using ARIMA (Autoregressive Integrated Moving Average) and Recurrent Neural Networks (RNN). Leveraging Firebase Firestore for cloud-based data storage and Chaquopy for Python-Java integration, the application enables users to input or upload historical sales data and generate 1-, 2-, and 3-month predictions. The app provides a user-friendly interface with visual representations of forecasts and supports algorithm selection between ARIMA and RNN. Empirical testing using real-world sales data demonstrates that ARIMA performs well on stable trends, while RNN offers advantages for fluctuating or nonlinear patterns. This project highlights the feasibility and benefits of embedding advanced predictive models into mobile platforms, enhancing accessibility for small-scale retail decision-makers.