Replication code for "Estimating the number of infections and the impact of non-pharmaceutical interventions on COVID-19 in 11 European countries" by researchers at Imperial College London. The paper about the model can be found here.
COVID-19 forecasting project, University of Oxford. One-year forecasts, country-by-country, based on a global model of travel patterns running on high-performance computers, that also try to incorporate mitigation measures and technological development in contract tracing and vaccines.
Pandemic Preparedness Planning for COVID-19-CovidSIM. “A deterministic simulation tool and planning tool for public health departments, local governments, companies and other parties” .
Forecasting COVID-19. Two scenario forecast, country-by-country, daily update.
A simple SEIR model (susceptible - exposed - infected - resistant) similar in scope to CovidSIM above. Not currently web-interactive, but downloadable as a runnable, editable parameters model for the Vensim simulation software, which has a free noncommercial version. Videos and a SIR (susceptible - infected - recovered) modelling tutorial are also available.
Modelling transmission with current assumptions. Team at the Big Data Institute, University of Oxford. For more information on team projects, visit coronavirus-fraser-group.org.
Epidemic curve taking a mitig into accusation strategy (many predictions do not do this). Source: https://github.com/sarusso/Covid-pred. Author note: this uses data from the 2003 SARS to extrapolate a fourth grade fit, on which I added noise and trained a LSTM neural net in hope to take into account for the mitigation. But it is not providing great results. It was more of a test.
Agent-based simulation model for the spread of COVID-19 in a healthcare district. It takes into account the differences in outcomes based on patient age, capacity of the healthcare system (hospital beds, ICU units), and non-pharmaceutical public health interventions (such as banning mass gatherings and limiting mobility). Implemented in Python.
Models of cases outside China by Rob Wibin.
Model of Confirmed cases, Deaths and Recoveries Globally. Uses the Gompertz growth model to forecast metrics by geo.
Forecast of Cases. Forecast of worldwide trends using multiplicative trend exponential smoothing model. Using Naive models might not be very reliable.
COVID-19 modeling. Currently exponential model with and without confidence intervals have been implemented.
SEIR model with no seasonal effect. High-level, very indicative modelling of true cases, new infections and deaths in any area based on few configurable parameters.
Paper (awaiting peer review) on role of climate in COVID-19 mitigation strategies.
This web application serves as a planning tool for COVID-19 outbreaks in communities across the world. It implements a simple SIR (Susceptible-Infected-Recovered) model with additional categories for individuals exposed to the virus that are not yet infectious, severely sick people in need of hospitalization, people in critical condition, and a fatal category.
Forecast of amount of infections, hospitalisations, ICU needed, respiratory systems needed and more for Belgium.
A discussion on models used by the UK government.
SEIR model with seasonal effect. Scientific Pandemic Influenza Group on Modelling (SPI-M) consensus view. The collection above is advice from academics to the UK government based on modelling.
🗣[email protected], [email protected], [email protected], [email protected]
Model of peak case numbers in the UK without intervention.
Feasibility of controling outbreaks of COVID-19 by isolation. Study by London School of Hygiene & Tropical Medicine.
A spatial model of COVID-19 transmission in England and Wales: early spread and peak timing. Data and code available at https://github.com/ldanon/MetaWards
Imperial College COVID-19 Response Team. MRC Centre for Global Infectious Disease Analysis, World Health Organisation Collaborating Centre for infectious disease modelling. (Report 9 - 16 March 2020) Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. "Perhaps our most significant conclusion is that mitigation is unlikely to be feasible without emergency surge capacity limits of the UK and US healthcare systems being exceeded many times over."
Critique of UK strategy by two public health schools in the US.
UK Short-term Deaths Forecasts (14-28 days) compared with Italy, Spain & France. Currently only uses exponential growth trends. Does not factor in any sigmoid behaviour so validity beyond mid-April is unlikely. Author: Dr Gareth Davies (Gruff).
Bret Victor-style interactive model of cases in London. Forecast of London cases using an exponential growth model. Users can change the key parameters using a UI built with Bret Victor's Tangle framework. Prioritizes intelligibility over accuracy.
Article (awaiting peer review) on the need for serological surveys to assess the stage of the epidemic.
🇺🇸🔗SF Bay Area CoVid-19 Model
Models of cases in San Francisco Bay Area by Max Henderson.
Short-term forecast of U.S cases. Estimating over 200,000 U.S. cases by April. Exponential fit, not logistic, but should hold in the short term.
Python open source short-term forecast of U.S cumulative cases. Logistic function fit. Daily updates with best scenario and worst scenario short-term forecasting diagrams.
Github project link is here: https://github.com/farawayboy/Covid-19-Prediction
A compartmental epidemiological model, based on the classic SEIR model, to describe the spread and clinical progression of COVID-19. Compares case load to healthcare capacity.
COVID-19 Italy hospitalisation and health estimates.
Critical care utilisation for the COVID-19 outbreak in Lombardy, Italy.
When does hospital capacity gets overwhelmed in Germany?
When does hospital capacity get overwhelmed in the USA?
When the US will run out of beds?
Ebola Map with different mitigation strategies.
Hospital capacity simulator/monitor used by Belgian health care. Allows to estimate impact of government measures on the course of the epidemic. Useful simulator to be launched soon.
Effect of travel restriction in Wuhan.
Effectiveness of contact tracing and isolation. “For R0 values of 2·5 or 3·5, if there were 40 initial cases, contact tracing and isolation were only potentially feasible when less than 1% of transmission occurred before symptom onset.”
Models forecasting the situation if stricter policies, similar to those in South Korea and China, were adapted in Europe and US.
Model of deliberate exposure by Robin Hanson.
Effect of Social Distancing on the Spread of COVID-19, primary research article.
Article. Potential Impact of Antiviral Drug Use during Influenza Pandemic
Consumer activity during COVID-19. Foot Traffic by Day Year-over-year foot traffic (aggregated by day) from SafeGraph.
CHIME, the COVID-19 Hospital Impact Model for Epidemics. A tool for COVID-19 capacity planning using discrete-time SIR modelling of infections/recovery.
R Epidemics Consortium (RECON). Includes a number of open source modelling tools and communities working on outbreak modelling.
Guesstimate. Simple modelling tool employing Monte Carlo simulations to account for random events.
Company using big data to provide early warning on the spread of infectious diseases.
Crowdsourcing (via social media) data on people losing their sense of smell in order to predict where people are infectious.
Make Your own COVID Dashboard with Google Sheets. Tutorial for the aspiring amateur data scientist: how to import a dataset and manipulate it to produce the charts you need to make sense of the pandemic.
Experiments with toy SIR models.
Daily worldwide figures on COVID-19 cases and deaths from the European Centre for Disease Prevention and Control. Excel, CSV, JSON, XML.
The Johns Hopkins daily US and worldwide figures on COVID-19. CSV.
Public and free to use. A machine readable dataset by people at Johns Hopkins University that aggregates relevant data from around ten governmental and academic sources on the county-level. The dataset contains more than 300 variables. A detailed description of them can be found here.
New York Times data on COVID-19 for US states and counties. CSV
Set of 29K articles (13K with full text) on COVID-19 inviting NLP (Natural Language Processing) analysis after "call to action" made by the US White House to technologists in March 2020. JSON
🇫🇷🔗https://www.data.gouv.fr/fr/datasets/donnees-des-urgences-hospitalieres-et-de-sos-medecins-relatives-a-lepidemie-de-covid-19/ & https://github.com/opencovid19-fr/data
French government datasets on COVID-19.
Italian data from government. The @DPCgov shared all latest #COVID19 data on @github: National trend - JSON data - Provinces data - Regions data - Summary cards - Areas
UK government data on COVID-19.
Round-up by the UK's Office for National Statistics (ONS) of what Coronavirus data they've published.
Coronavirus data for Germany.
🔗https://covid19.isciii.es/ & https://github.com/datadista/datasets/tree/master/COVID%2019
Coronavirus data for Spain. CSV files available (see bottom left for the first link).
Kaggle dataset challenge to build a model that helps predict the progression of the Coronavirus. Loads of kernels already uploaded: browse to “Kernels” or “Tasks”.
🗣Parker Ence: [email protected]
Veraset Data. Free dataset (for use for COVID-19 modelling only) of anonymous always-on GPS pings from smart phones. Hourly pings from roughly 1.5% of worldwide phones every hour (and pings from almost 10% of U.S. phones).
COVID-19 data and projections down to the US county level.
🗣 [email protected]
Metaculus. A forecasting community making predictions on different aspects of Coronavirus. Also running a forecasting competition.
Good Judgment Project. A forecasting community making predictions on Coronavirus spread.
Smarkets. A market on whether Coronavirus will be a WHO public health emergency on 6 May 2020.
🗣 [email protected]
National Coronavirus Preparedness Perception Tracker (US). Key measures of how Americans are feeling about the pandemic.Updated daily since March 12 2020, based on a daily national representative sample.
COVID-19 prediction (Colab Notebook). No author identified.
💡Better forecasting using Datasets
Data currently doesn’t correct for how good countries' testing regimes are. Could be combined with errors to allow for better predictions.
💡Capacity Limit Prediction
Medical Capacity Predictions for Facilities and Concepts to assist facilities with managing overwhelm.
💡Probabilistic model for estimating carriers
Since not every individual with symptoms is tested, there needs to be a way of estimating carriers en masse. A probabilistic model can help solve this problem. This idea is based on research by Augusta University into whether mass questionnaires of people at home could contribute to a prediction model, and help to distinguish high-risk cases from low-risk cases. "Population network structures, graph theory, algorithms to match subgraphs may lead to better clustering of households and communities in epidemiological studies,” Epidemiology and Infection (Cambridge University Press).
💡 https://www.defeatcovid.org/ (Help us by filling out the form too!)
Survey does not include name and exact address of individuals.
COVID-19 modelling tool in development needs volunteers now! We need volunteers with a variety of skills to help build / maintain an app to slow or stop the spread of COVID-19 using anonymous secure contact tracing and local evolving heat maps built with anonymized GPS data. Skills needed:
✉️ [email protected]
💡App : (Wireframes in development) Dataset from the survey will be utilised in the model to identify risky localities and hence city dwellers can be extra cautious while traversing via those localities. The device location history can also alert people to refill surveys if they have been in risky areas. Looking for individuals (code, research, data, web, app) who can help in developing this project.