meszdav / road-surface-quality

Udacity Machine Learning Engineer Nanodegree final capstone project.

README

Introduction

I live in Hungary, where the road surface quality is one of the worst in Europe. While I was on the road, I thought several times that it would be good to plan my trip depending on the road quality. In this case I could better enjoy the beauties of my country, and the trips would been more safe. Because of this, I decided to create a proof of concept in this topic.

The goal of the proof of concept is to evaluate specific road surface measurements, and plot the road surface quality to a geographical map, only using the measurement data.

Install dependencies:

v01

If you use pipenv:

Install pipenv via pip

pip install pipenv

After the pipenv installation. You need to run in the folder where you would like to create your virtual enviroment the following code:

pipenv install Pipfile

This command will install all the necessary dependencies. Now you are good to go.

v02

You can install all the necessary dependencies vie the virtualenv method. First you need to run the following command in a folder:

virtualenv env

After that activate your enviroment (on windows):

env\Scripts\activate.bat

Now run the following code:

pip install -r requirements.txt

Run the Project:

The jupyter notebooks are in the notebooks folder. In order to have all the necessary files for the machine learning modeles, you need to run the notebooks in this order:

1. rsq-data-processing.ipynb
2. rsq-feature-engineering.ipynb
3. rsq-ml-local.ipynb
4. rsq-dash-app.ipynb

All training and test data are in the data folder. The zip file in the data folder is the test measurement for the dash application.

The dash application can be executed via the rsq-dash-app.ipynb or via the dash-app folder with the runing of the application.py.