covid19-forecaster

Analysis toolkit for forecasting the revenue impact of COVID-19 on the City of Philadelphia's finances.

Read the reports:

• "Analysis of the Continued Impact of COVID-19 on the City of Philadelphia’s Tax Revenues" (March 2021)
• "Estimates of the Impact of COVID-19 on the City of Philadelphia’s Tax Revenues" (April 2020)

Installation

The following commands should be run from the command line, e.g., the Terminal app on MacOS or Command Prompt on Windows.

Step 1: Make sure you have the poetry tool installed.

Installation instructions are available on the poetry documentation.

Step 2: Clone the repository

From the command line:

# Clone the repository
git clone https://github.com/PhiladelphiaController/covid19-forecaster.git

# Change to the new folder
cd covid19-forecaster

Step 3: Install the dependencies

We will poetry to install the necessary dependencies into their own virtual environment.

From the covid19-forecaster folder, run:

poetry install

This will install all dependencies (as well as the development dependencies).

Running the forecasts

From the command line, the March 2021 forecast can be run using the following command:

poetry run covid19-forecaster results/

This will save data from the forecast to the results/ folder.

Working in an interactive environment

You can run the software in an interactive environment using Jupyter lab by running:

poetry run jupyter lab

This should launch a Jupyter window on your browser, at which point you can create a new notebook file (.ipynb file).

To run the software from the notebook, first import the March 2021 model (denoted as "v2"). From within a cell, run:

# Import the March 2021 version of the forecast (v2)
from covid19_forecaster.v2 import *

Now, you can examine the forecasts for individual taxes. For example, for the Wage Tax model:

# Load the wage tax model
wage = WageTaxForecast()

Examine the 'moderate' scenario:

wage.run_forecast("moderate")
wage.plot()

Examine the 'severe' scenario:

wage.run_forecast("severe")
wage.plot()