Inspiration
I’m an international student who moved to Canada to study computer science. When I first arrived, I experienced seasons for the very first time—and I had no idea what to wear! I didn’t even know that wind speed could affect how cold the weather feels. It was a struggle to figure out how to dress appropriately. That’s why I created this web application—to help others who might be in the same situation I was in.
What it does
A simple web application that recommends an outfit based on real-time weather data using a machine learning model.
How I built it
Built using Flask, Tailwind CSS, and scikit-learn. Python was used to train the ai models and was used to implement the api to gather weather data. The weather data was gathered according to the geolocation data: for this I used a geolocation data that is fetched via HTML5. When the user clicks on the button to get outfit, a POST request was sent to the flask backend, which loads a RandomForestClassifer to make a recommendation of what to wear.
Challenges we ran into
I struggled with making sure that the ai was accurate. I had to change the training model for the ai numerous times, as my model was being too vague. I struggled with the dataset as well, the dataset was too small and I had to build upon the dataset in order to get an accurate enough result according to the location.
Accomplishments that we're proud of
It works across numerous locations. Accomplished to build a new ai model from creativity - it appears to be a new idea that has not been thought of. The originality in the idea is something which I am very proud of. I believe, it can be useful to real life situations, as it saves time when someone wants to check what to wear for the day- they do not have to go through various steps in order to find out what to wear.
What we learned
I learned how ai models work in predicting data. Learning how to utilize Pandas to handle data, was something which I also dived into deeper; label encoding was something that was new to me as well; learning how to structure machine learning input from output and train/test split. After learning how to evaluate the model and generate a report, I had to utilize that in order to improve upon my model. RandomForestClassifier was something which was unheard of for me, before I dived into this project.
What's next for Outfit predictor
I have a google console account, and I plant to utilize the logic for this web application, in order to implement it as an app which can be downloaded from google play Store. I would also like to provide links to sites where they can buy the clothes for the weather if they do not have the clothes with them.
Log in or sign up for Devpost to join the conversation.