Uber Fares is a Data Science and Machine Learning I worked on in my free time. Its goals were to analyse the dataset of 200k NYC Uber rides and build a model to predict the price of the trip.
During the project development I have...
- Downloaded the dataset from Kaggle
- Formatted, cleaned, and enriched the dataset with additional data (NYC Neighborhoods and US Holidays)
- Created qualitative, spacial and temporal visualisations with Seaborn
- Iterated through several ML algorithms such as Polynomial regression, ElasticNet and Decision trees
Documentation is hosted on Netlify and built on Sphinx
├── data
│ ├── external <- Data from third party sources.
│ ├── interim <- Intermediate data that has been transformed.
│ ├── processed <- The final, canonical data sets for modeling.
│ └── raw <- The original, immutable data dump.
├── docs <- Sphinx Docs; see sphinx-doc.org for details
├── models <- Trained and serialized models
├── notebooks <- Jupyter notebooks for explorations
│ ├── 0.1_data_processing_tests
│ ├── 0.2_exploration
│ └── 0.3_machine_learning
├── references <- Data dictionaries, manuals, and all other explanatory materials.
├── reports <- Generated analysis as HTML, PDF, LaTeX, etc.
│ └── figures <- Generated graphics and figures to be used in reporting
├── utils <- Source code for all analysis
│ ├── data <- Scripts to preprocess data for analysis
│ ├── features <- Scripts to build features
│ ├── models <- Scripts to train models
│ └── visualization <- Scripts to produce visualisations
├── web <- Web demo
├── environment.yml <- Template for conda environment creation
├── Makefile <- Makefile with commands like `make data` or `make model`
├── pyproject.toml <- Python project config file
├── README.md <- The top-level README for developers using this project.
├── requirements.txt <- Pip requirements
├── test_environment.py <- Script for testing the correct environment setup
└── tox.ini <- tox file with settings for running tox; see tox.readthedocs.io