Ishan B's Senior Project Blog
 |
Project Title: AI Business Location Finder BASIS Advisor: Johnson Truong Internship Location: Independent Onsite Mentor: Krystian Confeiteiro, Student; Embry-Riddle |
Project Abstract
How can we use AI to make the job of picking a business's starting location easier? For a prospective business owner, picking a location can be a challenge. There are many factors that go into picking a location, and there is no one best location for every business. Each business has its own goals, and depending on these goals, the location will give different benefits, however a bad location can also lead to a business's failure A big reason that many businesses fail is because of factors such as insufficient market demand, ineffective business planning, and ignoring customer needs. For prospective business owners who know what type of business they want to start, picking a suitable location can help mitigate some of these challenges. The overall process for this research will be data collection: most of which will come from US census data, data preprocessing through libraries such as numpy and pandas, analyzing different models such as tensorflow and sklearn, model training, and then model tuning/validation. My goal for this project is to gain experience with a different aspect of AI then I already have experience with and learn about numerous different AI classification models. I also want to learn more about how to process datasets and prepare them for model training.
Week 10: Polynomial Regression
In my last post, I discussed the different metrics I used to analyze the performance of the linear regression (RMSE and score). I also discussed my belief for why the linear regression model performed poorly. This week, I trained the polynomial regression model. The metrics gave numbers that corresponded to my beliefs, but were still... Read More
Week 9: Model Adjustments
Last week, I faced some unexpected results with the linear regression model. The score of the model was exceptionally low. My main ideas were that either there were too many features so the model was having difficulties fitting to the data or that the data is too far from a linear trend to have any... Read More
Week 8: Model Analysis
Last week, I began training the sklearn model. The first sklearn model I trained was the linear regression model. When analyzing the usefulness/accuracy of a model, one of the simplest and first ways that it should be done is using the score method. The closer the score method is to 1, the more accurate the... Read More
Week 7: Model Training
Last week, I finished up the feature engineering generating features such as percent of population with different levels of education as well as comparisons of amounts of people with bachelor’s degrees to people with master’s degrees. This can help as certain business types are more likely to get customers of certain education levels. This week,... Read More
Week 6: Feature Implementation
Last week, I discussed feature engineering and how it works, specifically the correlation between different features such as people with "less than a high school diploma" compared to the different districts to see the amount of people with different degrees is directly related to the population or if there are other factors that go into... Read More
Week 5: Feature Engineering
Last week, I discussed the challenges I had with setting up the dataframe. The first challenge of changing the county names to numbers and assigning those numbers to the names was relatively easy. The other issue was a bit more of a challenge, but using inbuilt pandas methods, specifically iloc, I was able to loop... Read More
Week 4: Dataset Formatting
This week, I have been focusing on formatting the datasets from the US Census into the form that regression models use for their training. This is a challenging task as regression models need numerical data, and the county names are words. One strategy I may use to allow my program to still output the county... Read More
Week 3: Programming
Last week, my research focused on gaining a better understanding of regressions and how they work. This helped narrow my choices of which models to test to linear, polynomial, ridge, and lasso regression. After deciding which type of models to test, the next step in my research is to set up the programming interface. For... Read More
Week 2: Regressions
When making a choice of where to open a business, there is no correct answer. Each location has their own pros and cons, and the business owners preferences also factor into the decision. Because of this, I plan to use a regression model as that can help rank the different locations rather than just outputting... Read More
Week 1: Model Research
This week, I have been researching the different types of regression models to better understand which one will be optimal for my use case. Some examples of the different types of regression models are linear regression, logistic regression, polynomial regression, and lasso regression. After deciding which one I will use, I will find open source... Read More
Introductory Post
Hi everyone! I am Ishan Bansal and my senior project is called AI Business Location Finder. Small businesses contribute greatly to both the economy and community and can be classified by income, number of employees, or other metrics. For my project, a small business will be defined as a business projected to have less than... Read More