-
Our team logo
-
Home tab
-
Model page
-
Data Analysis page
-
Resources page
Inspiration
While brainstorming, we kept thinking about issues that Asian American women face and stumbled upon Lung cancer after reading a surprising statistic. Our team specifically wanted to focus on the Data and Social impact track. We are also aiming for the UI/UX, AI for education, and Women-centric tracks as well. We then decided to move forward with the idea while keeping in mind that we wanted to work on creating an ML prediction model using CT scan data from Kaggle.
What it does
We created a website showcasing our AI model that predicts whether a patient has lung cancer, based on their CT scan. The website also includes information and resources regarding lung cancer. The Information tab consists of histograms that showcase our analysis of a Lung cancer dataset that we utilized. The Resources tab consists of additional links and resources that our website users can visit to learn and increase their knowledge of Lung cancer stages, symptoms, prevention techniques, more different types of cancer, and etc.
How we built it
We used React and Node.js for the website. We used Python and TensorFlow for the AI model and used Python and Pandas for creating plots of the demographic data. We obtained our datasets from Kaggle.com using the following:
A survey dataset from lung cancer patients: https://www.kaggle.com/datasets/thedevastator/cancer-patients-and-air-pollution-a-new-link/data CT scans from the Iraq-Oncology Teaching Hospital/National Center for Cancer Diseases (Q_OTH/NCD) lung cancer dataset
Challenges we ran into
We had difficulties finding the right datasets for us, some datasets did not have a large enough sample size, some were missing data values, and so on. Another challenge that we ran into was with our ML prediction model and getting it to input onto our website. Project ideation and finalizing it took a while. A pro was figuring out how to utilize our different skill sets. It was very difficult to learn how to host a next.js project in Google cloud compute and our AI Prediction model kept crashing and not cooperating.
Accomplishments that we're proud of
Creating an image classification model that can recognize CT scans accurately. Creating a fully functional website in under 24 hours.
What we learned
We learned how to work in a team setting and realized that there is strength in having different skill sets that we brought to the team. Our teammates also learned how to further analyze data by analyzing data sets and determining what type of plots to create with the data we were able to find. Another thing we learned was how to create a ML prediction model using images of CT scans in Tensorflow. We also attended workshops and learned more about using Python and the pandas package.
What's next for Lung Lens
With the datasets we have, and how our model can allow for our datasets to evolve, we hope to put the dataset to use and create an educational CT scan classification test that doctors in training or anyone interested can practice their skills with.
After parsing through demographic data surrounding lung cancer, we hope to add more user interactivity by adding a survey that users can fill out to see what their risks are based off of that data.
Log in or sign up for Devpost to join the conversation.