😌 Get the Pet That's Best for Your Lifestyle

🧐 Problem Statement and Research Summary

If you're like any of us, you have probably never had a scorpion for a pet. But what if you would really like it? One of our team members had a tarantula for a pet, but she accidentally killed it. Anyway, the point is, you might not know what pet is best for you! This survey is supposed to help you figure out the best pet for your lifestyle. We didn't want to tell you what kind of pet to get because we're obviously not experts, so instead we asked people what kind of pet was best for them and some other information about themselves. Take this as just a recommendation for a pet that might be good for you. Remember that all pets require a lot of care, even tarantulas, which need live crickets, water, and a lot of soil to burrow in. Don't find out the hard way. Anyway, that's not what our project is about. The point is, take the survey to see what kind of pet might be best. Enjoy!

🎣 Data Collection Plan

The data for this model was collected on Amazon's Mechanical Turk! Mechanical Turk allows you to get a lot of survey responses quickly by paying people for their time. People can make a bit of extra money by completing short tasks while they ride the bus to work or wait for a friend, or even just to avoid boredom. You must be 18 years or older to be a worker.

💥 Ethical Considerations (Data Privacy and Bias)

Data privacy: we used a google form that was collected anonymously, since you honestly never know when someone will think some of their data is personal to them. Even though this is just a survey about pets, we tried to respect people's privacy. Also, it has been shown that you can identify a person based on just a few things about them like their date of birth, gender, and zip code. So it's just one more step for people to be figured out based on their pet preferences and how much free time they have. We just wanted to make sure that people who were giving us their time to take this survey felt safe.
Bias: we were only able to collect about 95 or so responses in our survey, so our data is pretty limited. You can tell by the accuracy of the model that we weren't able to get enough data because it's not very accurate. This could cause bias because the model might not reflect the real world. What if actually people really want to adopt scorpions, just not the people we surveyed? We may never know tbh. However, people aren't using our data to make life-changing decisions, probably, so just keep informed about the possibility of bias and you won't base a big decision on a model that isn't perfect. Our project can be kind of like a suggestion for you.

👻 Our Team

This app was designed and built by The Curriculum Team at The AI Education Project.