This is a web user fronted software that lets the user calculate an estimated price of their motorbikes from home, letting them have a preview before starting contact with any agent, or selling it for a worse price.
This project makes the calculation based on a Machine Learning System that takes data from different motorbikes (previously extracted from second-hand motorbike online shops with our tool Motonitor-Deep-Scrapper) and uses it to make a learning model that will predict with a pretty good accuracy (82% scored) an estimated price for the motorcycle.
First, we need to download python 3.9.2
or higher is required, and preferably a python3-venv enviroment to install all the required packages.
If you are using venv you can easily create a new enviroment with "python3 -m venv venv" (Inside project's folder).
Activate the python3-venv with "source venv/bin/activate"
Once you have your enviroment setted up and ready to work with python you need to install the requirements:
pip install -r requirements.txt
Now that we have already installed all the required packages we are going to start it:
./manage.py runserver
Now you can acces the user friendly website using this url http://127.0.0.1:8000/ (or the specified port).
Just the user friendly front-end.
The backend, where the magic happens.
The machine learning model with the preprocess, train and split, learn and predict functions that makes our software a reality.
The scrapped data used to train the ML-ModelS.
A little API to connect with our front-end.
A Jupyter Notebook used for internal purposes on calculating the best scores for our Machine Learning Models and analyzing them with graphics.
In the development of our software we did some analysis of the data that we were getting from our Machine Learning Models.
Our score calculation results:
Motonitor's linear regression line:
Motonitor
is free and open-source software non-licensed. Official Motonitor was created by Jorge Vinagre Triguero, Zhensheng Chen, Roberto Lupu and Joel Otero Martín for the HackUPC 2022 BarcelonaTech student Hackathon in 36h in a non-stop and non-sleeping rush submitted here.