Abstract
This Independent Study aims to explore the use of Markov chain models for predicting population and migration patterns between states in the United States. This is done by using the 2019 population and state-to-state migration flow dataset from the United States Census Bureau. We develop Markov chain models to predict the population and probability of moving from one state to another for any given year and use these results to answer three main research questions:
In a regular Markov chain model, the probability distribution of the states often reaches an equilibrium at some time. Based on the population forecast, when do we expect the state-to-state migration probability to reach its equilibrium? Does the ranking of state by population remain constant until it reaches equilibrium?
Based on the Markov chain population projection, which states have experienced the largest inflows and outflows of migrants? Does the prediction of the number of people entering and leaving each state remain constant or change over time?
How accurate is the population forecast generated using Markov chain models? Assuming that it is not completely accurate, how far off is the forecast from the official projection done by the US Census Bureau and how much does the error vary over time?
Our study finds that the convergence time is 132 years, meaning that in the year 2151, all states are predicted to have a similar population of 14,795,131. Initially, California, Texas, Nevada, Oregon, and Hawaii were the states with the largest migration inflow and the District of Columbia, Delaware, Alaska, Wyoming, and Rhode Island experienced the largest migration inflow. Due to different rates of increase and decrease in population for each state, the ranking of state by population changes over time. We also found that although Markov chain models do not accurately predict population changes over time, our predictions are comparable to the actual population, where an increase in the predicted population indicates an increase in the actual population and vice versa. These results are presented on an interactive ReactJS website.
Advisor
Guarnera, Heather
Second Advisor
Pierce, Pamela
Department
Computer Science; Mathematics
Recommended Citation
Ziku, Rekik, "Visualizing Movement: Predicting Interstate Migration Probabilities and Populations of U.S. States Using Markov Chain Analysis" (2023). Senior Independent Study Theses. Paper 10465.
https://openworks.wooster.edu/independentstudy/10465
Publication Date
2023
Degree Granted
Bachelor of Arts
Document Type
Senior Independent Study Thesis
© Copyright 2023 Rekik Ziku