Chicago COVID-19 Forecaster

Chicago COVID-19 Forecaster is a forecaster program that takes data from google search trends, Chicago street congestion data, and divvy bike data to formulate a forecast on the spread of COVID-19 in Chicago.

Uses

Any code and/or data set can be used under the MIT license. If using this repository, please be sure to credit it.

Python Scripts

• datagather.py, Gathers data from multiple APIs and puts them in the master data set, trend.csv
• forecaster.py, Machine learning script that makes a model to forecast new cases for the next 3 days. This script makes a 14-day data set with 11 real days and 3 predicted days in, prediction.csv
• graphic.py, Makes a graphic to show the 3 predicted case days along with positivity rate and new case graphs. This graphic is used for the @ChiCovid19 twitter.
• makesets.py, Makes many different data sets to train our machine learning model and to predict new cases for the next 3 days.
• visualizedata.py, Visualizes the 7-day moving average of the raw data (this data set is avg.csv).

Data CSVs

1. trend.csv, raw data values in with columns (date, cases, gtrend, traffic, bike)

• date, YYYY-MM-DD
• cases, confirmed cases
• gtrend, google trend average score for "covid symptoms" and "covid testing near me" in Chicago
• traffic, average GPS pings in different Chicago regions
• bike, average bikes being used per dock in Chicago

2. avg.csv, all raw data values put into a 7-day moving average, first 6 raw days are removed.
3. prediction.csv, 14-day data set with 11 real days and 3 predicted days
4. shifted3.csv, all columns except for date and cases shifted down 3 days
5. avgshifted3.csv, all columns except for date and cases shifted down 3 days, columns are then set to a 7-day moving average
6. nobikeshifted3.csv, bike column removed, all columns except for date and cases shifted down 3 days
7. avgnobikeshifted3.csv, bike column removed, all columns except for date and cases shifted down 3 days, columns are then set to a 7-day moving average

Order Of Scripts

• Scripts should be run in the following order to ensure proper results

1. datagather.py
2. makesets.py
3. forecaster.py

• The following scripts have no proper order and serve as visual aids

1. graphic.py
2. visualizedata.py