Short Stack Sustain Machine Learning Talk

Getting Started

General Prerequisites

NOTE: The suggested code instructions are for macOS

Install Visual Studio Code

Install Visual Studio Code here: https://code.visualstudio.com/

Add VS Code to your PATH

Follow instructions here https://www.freecodecamp.org/news/how-to-open-visual-studio-code-from-your-terminal/

Installing brew and necessary packages

Use the brew package manager to install requirements before setting up the repo. Visit https://brew.sh for installation instructions for the package manager and then enter following command to install requirements.

brew install [email protected]

Installing python through pyenv

Installation of Python can also be done through pyenv which can be installed through brew.

brew install pyenv

# Make sure to update x with the correct version of 3.10 in use. You may list all versions of 3.10 by executing
# pyenv install 3.10

pyenv install 3.10.x
pyenv global 3.10.x
pyenv user 3.10.x

If you are using zsh, execute the following to make sure the python command will link properly to the version installed through pyenv:

echo 'eval "$(pyenv init --path)"' >> ~/.zprofile

echo 'eval "$(pyenv init -)"' >> ~/.zshrc

Instructions for Setup

Clone the Repository

git clone [email protected]:connor-hilll/short_stack_ml.git

Setting up via CLI

Follow these instructions and run the commands from the root of the repo.

1. Create Virtualenv:

   • python3 -m venv venv

2. Activate venv:

   • . venv/bin/activate

3. Install requirements:

   • pip3 install -r requirements.txt

4. Install new kernal:

   • ipython kernel install --user --name=short_stack_ml

5. Open in VS Code:

   • code .
     • Note: VS Code must be added to your path. Reference Add VS Code to your PATH section above

6. Ensure iPython Notebook is using the virtual environment:

   • In the energy_analysis.ipynb file, in the top right corner, it should state that the kernal is venv (Python 3.x.x)

Notes

Packages

The main packages used are as follows:

1. Pandas - data analysis and manipulation tool
2. Numpy - fundamental package for scientific computing
3. StatsModels - provides classes and functions for the estimation of many different statistical models
4. Scikit-learn - Simple and efficient tools for predictive data analysis
5. Matplot lib - Plotting library
6. iPython - Provides a rich toolkit to help you make the most of using Python interactively

Some of the librarys have overlap such as scikit-learn and statsmodels. However, they all serve specific purposes

Data Sets

Weather Data

Can be found in ./data/weather_data.csv

NOTE: We are specifically interested in the Philadelphia Airport data

The weather data is free from NOAA (National Centers for Environmental Information) and can be found here: https://www.ncei.noaa.gov/

Energy Data

Can be found in ./data/residential_usage.csv

This is energy data from a single family residential property in Southern New Jersey. Unfortunately it was quite difficult to find free anonymous multifamily data.

The energy data can be found here: https://data.mendeley.com/datasets/rfnp2d3kjp

Cleaned Data

The cleaned data is a combined version of the two datasets. CDD and HDD stand for Cooling Degree Days and Heating Degree Days respectively. These values are commonly used in energy calculations. Please reference this weather.gov link for more information.