Welcome to UFO_predictor 🛸, an innovative and fun project developed during the LeWagon Lisbon data science bootcamp. Our tool allows users to predict and visualize UFO and alien visitations using advanced AI technologies.
UFO_predictor 🛸 was pitched by me and as team leader developed the project in a team of 5. The primary goal is to predict potential UFO sightings in the USA and provide visual representations based on eyewitness accounts using cutting-edge machine learning and generative AI models.
- UFO Sightings Prediction 🌍: Predicts the location, shape, and duration of potential UFO sightings using machine learning models.
- Generative AI Visualization 🤖: Generates realistic images of UFOs and aliens based on the predicted data and user inputs.
-
Machine Learning Models 📊: Our models predict UFO sightings by analyzing historical data on sightings. We utilize models such as:
- Location Prediction Model 📍: Uses clustering and regression techniques to predict potential sighting locations.
- Shape Prediction Model 🔺: Employs classification algorithms to determine the probable shape of a UFO based on past reports.
- Duration Prediction Model ⏳: Estimates the likely duration of a sighting using time-series analysis.
-
Generative AI Models 🎨:
- We used a combination of GANs (Generative Adversarial Networks) for UFO image generation 🛸 and OpenAI's DALL·E model to generate images of aliens 👽. The alien images are created by sending descriptive prompts to the OpenAI API, which returns a unique image based on the description. This approach provides a visual context to our predictions and adds an element of creativity and realism to the project.
You can explore our UFO Predictor app here: UFO Predictor App.
Simply enter the required data, and the app will provide predictions and visualizations based on the input. The user interface is intuitive and designed for easy navigation.
Our application is built using Streamlit, a powerful framework for building and deploying data science applications with ease. The app is also deployed on Streamlit, ensuring a smooth and interactive user experience.
For a detailed overview of the project, including data analysis, model development, and implementation, check our project PowerPoint presentation here: UFO Predictor Presentation.