🚴‍♂️ Dash Application for Bike Station Data Analysis and Prediction 📊

Welcome to our Dash application for analyzing and predicting bike station data in the city of Toulouse. Hosted on Google Cloud, this interactive app allows you to visualize data and predictions using various machine learning algorithms. The application is in French.

🔗 Link to the application: Dash App on Google Cloud

📋 Description

This Dash application provides comprehensive tools for data analysis and machine learning predictions on the bike station data from Toulouse. We have data on the percentage of bikes present at each of the 185 bike stations.

🌟 Features

For Data Analysis:

• 📈 Interactive Statistics Description Graph for each bike station
• 📊 Interactive Correlation Matrix
• 🗺️ Interactive Correlation Map
• 🧬 PCA Analysis
• 🔄 PCA Reconstruction
• 🔍 Eigen Vectors Analysis

For Predictions:

• 🧠 6 Prediction Methods:
  • Prediction by Mean
  • Prediction by PCA Reconstruction
  • Multiple Linear Regression
  • Random Forest
  • XGBoost
  • XGBoost with PCA
• 📊 Metrics Evaluations (MAE & MSE)
• 🌍 Interactive Map to visualize metrics
• 📈 Global Metrics

🛠️ Technologies Used

• Dash: For building the interactive web application.
• Plotly: For creating interactive graphs.
• Google Cloud: Hosting the application.
• Machine Learning Algorithms: Random Forest, XGBoost, Multiple Linear Regression, PCA.

🌐 How to Access

You can access the application via the following link: Dash App on Google Cloud

📚 Application Language

The application is in French. Please navigate through the features using the French interface.

📥 Installation

If you wish to run the application locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/Hisqkq/Time-series-prediction

2. Go to the app directory:

   cd DeployWeb/src

3. Install the required packages:

   pip install -r requirements.txt

4. Run the app:

   python app.py

📊 Visualizations

The application includes various visualizations for data analysis and prediction evaluation:

• Interactive statistics graphs for each bike station.
• Correlation matrix and map.
• PCA analysis and reconstruction graphs.
• Eigen vectors analysis.
• Prediction visualized on interactive line charts.
• Prediction metrics visualized on an interactive map and in global metrics graphs.

Enjoy exploring the bike station data with our interactive Dash application! 🚴‍♀️📊