PHS COVID-19 : Preparing for winter

Domain knowledge and the business context

Public Health Scotland is Scotland's national health improvement agency and one of the country's main national public health bodies. We take a population focus, working to understand and influence what makes communities, regions, and countries more or less healthy. These factors that impact on health are known as the determinants of health.

COVID-19 is a serious pandemic, which created a wealth of public health data. The focus of this analysis is to prepare for winter with COVID-19 in place using the data provided by the Public Health of Scotland.

Tools Used: R, Plotly, Slidy Presentation

Concepts used: Time Series Forecasting using ARIMA Model

Business Questions:

• How the daily positive cases changes over time ?

• What has the vaccination coverage and its progression been?

• Analyzing hospitalization and fatality trends

• Projection of hospitalization and deaths in near future

Data Source:

All covid data is provided by Public Health Scotland: https://www.opendata.nhs.scot/dataset/covid-19-in-scotland

Challenges:

The data was converted to Time Series for forecasting and model building. Handling and converting data to time series and incorporating them into model is quite tricky. Also the time series data uses autoplot for plotting which has its own restrictions when plotting. I converted them into ggplot and implemented ggplotly for interactive version

Functions:

A function calculate_roll_avg is created to handle the RollingAverage which is being widely used to calculate the increase or decrease in the cases.

A function convert_to_timeseries is created to convert the data to timeseries.

Analysis:

All the analysis documents are available in the analysis folder (both Rmd and R)

The overall analysis is available in the document analysis/Rmd/PHS_COVID_ALL.Rmd

The detailed analysis of individual topic is available in the following sections

1_Exploratory_analysis.Rmd

2_Vaccination Analysis (ARIMA Modelling).Rmd

3_Hospitalisation Analysis (ARIMA Modelling).Rmd

4_COVID_Deaths Analysis (ARIMA Modelling).Rmd

The files are labelled accordingly to the order of execution. The same files are also available in R.

Some plots demonstrating the insights

Daily Cases:

• There were total of 571,421 Positive cases reported as of 6th of October in Scotland.

• Between 23 September 2021 and 30 September 2021, 21,279 people had a confirmed positive test result. This shows a decrease of 9.66% compared to the previous 7 days.

Vaccination

As of 30th September in Scotland,

4,195,393 people had been given first dose 3,842,053 people had been given second dose

COVID-19 Hospitalizations and Deaths Hospital admissions

• There were total of 34,624 patients admitted in hospital and total of 2572 patients admitted in ICU.

• Between 23 September 2021 and 30 September 2021, 719 had been hospitalized with COVID Positive. This shows a decrease of 15.11% compared to the previous 7 days.

• There were total of 8726 Deaths reported.

COVID-19 Hospitalisation Projection: ARIMA Model(p,d,q):

COVID-19 Deaths Projection ARIMA Model(p,d,q):

Summary:

• Daily Cases trend seems to be declining as we see a decrease of 9.66% when compared to previous 7 days.

• Almost 91.43% of the people are vaccinated with Dose 1 and 84.45% of the people are vaccinated with Dose2.

Projections:

• Hospitalization shows a decrease of 15.11% when compared to previous 7 days.

• The ARIMA model projects that the hospitalizations might drop, based on the declining trend of the current COVID-19 Hospitalizations,

• The ARIMA model also projects that the weekly deaths maintains at the same level.

• However potential future changes in hospital occupancy and intensive care use is highly uncertain as it is also dependent on both current infection levels and how the vaccination impact will be.

Future Proposal:

• Even though the trends on daily positive cases, hospitalizations and death seems to be declining in Scotland, England shows an increase in all the categories.

• Given that the border is fully open, this model could be expanded to include data from the whole of UK which might give a different perspective.

• As a preventive measure, the travel restriction might help to reduce the spread of infection and control the hospitalization.

• With winter approaching, considering the previous trend, its better for the hospitals to be prepared.