Methods blog · Global Health & Epidemiology

COVIDSpread

COVIDSpread: real-time, customisable time-series forecasting of COVID-19 spread

A unified platform that couples live COVID-19 data visualisation with non-linear time-series models that users can fit to any window of the data, for short-term forecasts.

#COVID-19#forecasting#time series#epidemiology#R Shiny

During the pandemic, dashboards visualised COVID-19 case counts everywhere, but visualisation is not prediction. Very few tools let people move from “what happened” to “what is likely to happen next week” using the same real-time data. COVIDSpread, developed with Sahar Shahriari, Taha Hossein Rashidi and Fatemeh Vafaee at UNSW and published in F1000Research, fills that gap.

Design goals

  1. Real-time: work from the latest available data.
  2. Statistical, not just visual: fit forecasting models within the same platform used to explore the data.
  3. Customisable: let users decide which part of the epidemic curve to model, for example only the period after a lockdown began, because the dynamics change with interventions.
  4. Accessible: usable by lay users and by experts.

Models

Epidemic curves are rarely linear: growth accelerates, saturates and responds to interventions. COVIDSpread therefore relies on time-series models able to capture non-linearity. A generic short-term forecast takes the form

\[\hat{y}_{t+h} \;=\; g\big(y_t, y_{t-1}, \dots, y_{t-p};\, \hat\theta\big), \qquad h = 1,\dots,H,\]

where \(y_t\) is the (possibly transformed) count on day \(t\), \(g\) is the fitted model and \(H\) is a short horizon. Users can develop several customised models with different restrictions, such as a model fitted only to a chosen time window, and compare their forecasts side by side.

Why time windows matter

Fitting one model to the whole pandemic blends regimes: pre-intervention growth, lockdown decline and reopening waves. Restricting the window to the current regime gives forecasts that reflect present dynamics. Comparing windows is also a quick way to assess the effect of an intervention, which was the focus of our earlier medRxiv preprint, Real-time time-series modelling for prediction of COVID-19 spread and intervention assessment (2020).

Access

The platform (originally released as “COVID-19 TS”) was hosted by the Vafaee Lab at UNSW and is no longer online. The methodology and its evaluation are described in the open-access paper below.

Read the paper

  1. COVIDSpread: real-time prediction of COVID-19 spread based on time-series modelling

    Siroos Shahriari, Taha Hossein Rashidi, AKM Azad, Fatemeh Vafaee

    F1000Research 10, 1110 · 2021

    Q1 OA 8

BibTeX