Currently, nearly 35% of the world lives in poverty and around 689 million people live in extreme poverty. Several countries have resources to help these people but cannot always accurately identify where the resources need to be sent. Rather than using the expensive and biased method of survey collection, the researchers chose satellite imaging as it is cheap, available, and without bias. To identify regions affected by poverty, the researchers developed a logistic regression model and tested different features and hyperparameters to determine the best model; other linear classification models such as Support Vector Machines were evaluated as well. None of the features or models truly worked better than the others. Evaluating results on the accuracy, precision, and recall, the best accuracy achieved was 60%. Although there was no best feature or hyperparameters to optimize, the researchers found that the data was not linearly separable. (Summer 2021)