Note: This project is part of the Group Final Project (GFP) for the RevoU Fullstack Data Analytics Program's final task.
We are using US Regional sales as the dataset
After doing EDA and identifying the business problem, we found out that we could expand to state that we dont have store on, so we’re going to use Regression to find the city of a state that we can expand our business.
First, we gather data on the cities where our stores are located. Then, we select features (variables) that we can use to train our data. Because the number of cities is relatively small for use in machine learning, we decided to use LightGBM. It provides a decent score for data with a small number of rows.
The idea behind regression is to use X variables to predict Y. In this case, the X variables are demographic and geographic data used to predict total revenue (The Y). These X variables are what we call features.
After training and testing our data (using 300 rows for training and 62 rows for testing), we achieved a score of 0.699 for our regression analysis. We thought this was a pretty decent result.
Next, we collected data on the largest city from each state that we did not previously have in our store locations, along with its features. Here is what we found:
After using the models that we trained before the result are like this:
Our models suggest that the city of Portland, located in Maine state, has the biggest potential revenue. Therefore, we recommend expanding our business to that city.
Thank you for taking the time to review our project. We hope that you found it informative and insightful. As we strive for continuous improvement, we welcome any feedback or suggestions you may have. Your input is valuable to us, and we are committed to incorporating it into our work.
If you're interested in reviewing the notebook we used, you can find it at the following link:
If you have any questions or comments, please feel free to reach out to me. I am always available to discuss our project in more detail or provide further clarification on any aspect of it.