ConsumeUrDoom

Inspiration

After browsing a variety of datasets, we ran into a wonderful set containing a classification of characteristics of poisonous and edible mushrooms, which reminded us that it is mushroom picking season!

What it does

ConsumeUrDoom is a website that puts you in the setting of someone who is hungry and stranded in the middle of a hungry forest. The website explains that in front of you lies a mushroom, and invites you to describe it based on its cap colour, cap size, etc. Then, it offers you the choice to eat it, or discard it, and ends with the display of your fate.

How we built it

First, we designed the Machine Learning and used a Decision Tree Classifier as a model to determine whether the inputted mushroom is poisonous or edible. We used Velo by Wix for the aesthetics and used Domain.com to obtain a domain for our website. We used Flask to establish a connection between our front end (Velo by Wix) and our back end (Python Machine Learning).

Challenges we ran into

We had trouble finding a dataset that was easy to work with and didn't require any NLP (Natural Language Processing) to treat, but after 2 hours of searching, we finally found the dataset that suited our needs. Another problem we ran into, was understanding how to connect Flask to Velo by Wix, fortunately, we had help from the MLH representative.

Accomplishments that we're proud of

As Machine Learning is a concept that none of us have worked with previously, we were extremely proud of ourselves for successfully creating a functional model which can predict if a mushroom is poisonous or not based on its description. Using Flask and Velo by Wix was also another accomplishment, as we spent a good amount of our time figuring out how they both worked alongside one another.

What we learned

We learned that there are many different branches of Machine Learning, such as classification, which can be done with a decision tree or with logistic regression. The possibilities are endless! As mentioned previously, we had never worked with Flask nor Velo by Wix, so we became more familiar with both tools.

What's next for ConsumeUrDoom

One feature we would like to add is a visual representation of the mushroom you're describing. With each new question answered, the mushroom would look more detailed and more accurate to what you're imagining. This would be a cool feature, as it would allow the user to see what mushroom they would be agreeing to eat, or discarding. This could open the door to a more surprising aspect, as the worst-looking mushroom could still be edible, and the most harmless one could be poisonous!

Note: When connecting the Velo by Wix front end with the Flask back end, we encountered an error, which we didn't have time to debug. Despite that, our Machine Learning portion is still functional and the code can be tested manually.

Error mentioned above:

Access to fetch at 'https://yuanlu.pythonanywhere.com/predict?ignorethis=true${urlParams
}' from origin 'https://naziachowdhury064.wixsite.com
' has been blocked by CORS policy: No 'Access-Control-Allow-Origin' header is present on the requested resource. If an opaque response serves your needs, set the request's mode to 'no-cors' to fetch the resource with CORS disabled.