1) I believe that my social distance detector was definitely a lot more effective than I thought it would be. Although it captured many social distancing violations, it seemed that it only worked relatively well for people that were close to the camera lens. In the video I chose, there were a group of people in clear sight of the camera which the detector was able to distinuigsh pretty well, however, there was a large group of people in the backround that were unable to be distinguished by the detector. There were also a couple of times where the detector did not recognize an individual as a person. However, this makes sense because since the video I chose was a cheer competition, during some stunts when the competitors were higher up in the air, the detector could not recognize them.

2) Although the detector was effective at some points, I do not think it should be the main source for detecting social distancing violations. Using it in conjunction with other methods to detect violations could be very beneficial. However, if the camera angle is closer to 90 defgrees, the results can be far more accurate, making the detecor a good choice.

3) The first improvement, as I stated above, includes filming the people from a 90 degree angle so that it is easier to tell if they are 6 feet apart. With different angles, its hard to see exaclty how far apart individuals are. Furthermore, by increasing the testing set (testing a wider range of people) it is possible to get a more accurate values. For example, instead of just training the model on people walking, the training set could also include individuals riding their bikes. By training the model on more people, it is possible for the model to identify a wider range of individauls and more accuratley distinguish social distancing violations.