Predicting Bicycle Traffic in Zurich

Note

This project repository was part of a data science project at University of Zurich in 2024. It's content is now archived.

Image Source: www.zuerich.com

A data science project for the following lecture:

• University : University of Zurich
• Lecture: ESC 403 Introduction to Data Science

Methods used for predicting bicycle traffic in Zurich City:

• Random Forest (Python)
• Linear Regression (R)
• Lasso Regression (R)
• Poisson Regression (R)
• Negative Binomial Regression (R)

Authors

• Julie Tschanz
• Philipp Wyss (philippchristian.wyss@uzh.ch)
• Damian Brülhart
• Mike Krähenbühl

Project Proposal

Where do I find the project proposal?

📁 esc403
┗━📁 doc 
  ┗━📁 01_proposal
    ┗━📜 proposal_traffic_zurich.pdf # see here

Project Presentation

Where do I find the final project presentation?

📁 esc403
┗━📁 doc 
  ┗━📁 02_presentation # reproducable 🎉 Quarto presentation

Data Sources

• Stadt Zurich
• opendata.swiss (Swiss Open Government data)

What about data preperation?

We have created a seperate data preperation notebook for each year. That happens there:

1. Load raw data from 📁 data folder or directly from source
2. Aggregate data on hourly basis
3. Imputation of missing values
4. Joining of different datasets
5. Export as .csv to 📁 results folder

📁 esc403
┗━📁 src 
  ┣━📜 data_preparation_2019.ipynb
  ┣━📜 data_preparation_2020.ipynb
  ┣━📜 data_preparation_2021.ipynb
  ┣━📜 data_preparation_2022.ipynb
  ┗━📜 data_preparation_2023.ipynb

General Project Structure

General project structure is derived by "Good enough practices in scientific computing" - G. Wilson et al.

📁 my_project
|--📁 doc  # text associated documents and final hand-in's
|--📁 data  # raw data files
|--📁 results # files generated during cleanup and analysis
|--📁 src # project source code / functions / reports / dashboards
|--📁 bin # external scripts or compiled programs

Branching Strategy (Git)

1. Develop in your personal "development" branch
2. Merge to main when ready and tested

---
title: ESC403 branching strategy
---
gitGraph
   commit id: "initial commit"
   branch philipp
   checkout philipp
   commit
   commit
   checkout main
   merge philipp
   commit id: "branch off"
   branch julie
   checkout julie
   commit
   commit

Loading