Abstract
This Demographic Web Application is an undergraduate level project developed as part of the academic curriculum in the Department of Electrical and Electronic Engineering at University of Rajshahi. The project aims to provide a user-friendly platform for projecting the total population of Bangladesh based on user inputs. The application utilizes Next.js, React, and various libraries to offer interactive visualization and population trend analysis.
Acknowledgment
We would like to express our sincere gratitude to our supervisor, Prof. Dr. Mamun - Ur - Rashid Khandker sir for his valuable guidance and support throughout the project. His expertise and encouragement have been instrumental in the successful completion of this endeavor.
Algorithm
The algorithm initiates its process by gathering essential data from reputable sources, which includes death rates and infant mortality rates covering the extensive period from 2011 to 2101. These rates, averaged every five years, are meticulously collected from MacroTrends. Additionally, the algorithm incorporates suicide rates spanning the period of 2011 to 2023 from the same source. Crucial insights are drawn from total population figures and population distribution across various age groups, such as 0-4, 5-9, sourced from Wikipedia. A significant foundation for geographic understanding is established through the district and division data of the 2011 Bangladesh census.
Central to the algorithm's operation is the simulation of population dynamics across time. The initial step involves the division of the population of each district into specific age groups based on the percentage distribution of the total population within those groups. Key age segments, notably the 20-24 age group of each that represents potential new couples, are identified as pivotal contributors to future population growth.
When a user specifies a target year within the range of 2012 to 2101 and indicates the desired number of children per couple, the algorithm embarks on a series of iterative steps. These iterations span from 2011 to the user-specified target year, allowing for a comprehensive view of population changes. For instance, if the user selects the year 2101 as the target and inputs a desired average of 2 children per couple, the algorithm iterates through every five-year interval.
During each iteration, the algorithm calculates trends, estimates child populations by multiplying the population of the 20-24 age group (half of the group) by the user-provided number of children. Adjustments are made based on infant mortality rates, and the child populations are further refined. Moreover, the influence of death rates and migration rates on the overall population is factored in. To ensure a realistic model, the algorithm facilitates shifts of individuals between adjacent age groups as the years progress. For instance, individuals within the 15-19 age group transition to the 20-24 age group, while newly born children are included in the 0-4 age group. Suicide and migration rates are considered for the 15-19 age group. This meticulous process ensures that age distributions remain consistent and coherent throughout the iterative calculations. The algorithm's predictive power is rooted in its ability to continuously update population figures and age group distributions over the projected timeline.
Future Scope
The project forms the basis for prospective advancements and exploration in population projections. While the algorithm currently shifts age group populations, it doesn't incorporate age-specific death rates due to the unavailability of data. Future improvements may involve integrating age-specific death rates, provided data is attainable. Also we assume that individuals aged 20-24 constitute new couples, although data accuracy may vary. As we lack specific information on newly married couples, this assumption is made. Future improvements could encompass the inclusion of accurate data on new couples, provided such data becomes obtainable. Additionally, the application could delve into the realm of machine learning, and expanding the application's scope to cover other regions or countries.
Supervisor
Prof. Dr. Mamun - Ur - Rashid Khandker,Department of EEE,University of Rajshahi
Student