README.md

In R4IDSR/covidmonitor: Merge and Analyse WHO AFRO Covid-19 Monitoring and Evaluation Data

covidmonitor

The goal of {covidmonitor} is to provide functions for simplifying data merging and cleaning for WHO AFRO COVID-19 linelists and monitoring & evaluation data. The package further contains templates for automating report production.

Installation

Currently the package is not on CRAN. Once it is - you can install the released version of epichecks from CRAN with:

# install.packages("covidmonitor")

In order to install the package you will first need to install an extra bit of software called Rtools. You can download the installer from: https://cran.r-project.org/bin/windows/Rtools/ Please install the version highlighted in green.

Once this is installed and you have restarted your computer, the development version of the package can be installed from GitHub with:

# install.packages("remotes")
remotes::install_github("R4IDSR/covidmonitor")

If you do not have a LaTeX editor installed on your computer, then please run the following code to install TinyTex. This is needed in order to be able to create Word and PDF documents using R markdown.

install.packages('tinytex')
tinytex::install_tinytex()
# to uninstall TinyTeX, run tinytex::uninstall_tinytex() 

Country linelists

Folder structure

Functions and templates are set up to run based on files contained in the AFRO WHE HIM - Country LineList sharepoint folder. This folder contains all the latest country linelist files which are automatically sorted from files emailed to [e-mail] by countries, using a .net script.

To ensure that the data pipeline works in an automated fashion, please always email the country files to the email address specified. This way the latest file can be efficiently identified.

File naming and format

In order for the R functions to run, the files must be named appropriately, with the start of the file name containing the three letter ISO code for the country followed by a dot (.) and then the date of report; after this the rest of the name can be anything. For example a file submitted by Namibia on the 29th of January 2021 should have “NAM.2021-01-29” at the beginning of the name, and so the full file name could be “NAM.2021-01-29.Linelist.Namibia COVID-19 Confirmed.xlsx”.

All files must be stored as an excel file (i.e. xlsx), those stored as binary workbooks (xlsb) will throw an error and must be re-saved as xlsx.

n.b some countries (e.g. TCD and SEN) have saved their date variables in excel using the 1904 date system (i.e. date origin is 1st January 1904 rather than 1900), this needs to be restructured manually in the excel file otherwise the R function will produce dates that are off by four years. See here for details. Lines 484-486 can be uncommented in linelist_clean.R raw code to crudely deal with this issue but it is recommened to fix the issue manually and not use this quick fix.

Dictionaries

In order for the functions to work, several data dictionaries are required. The first is called linelist_dictionary.xlsx and contains information for defining and naming variables, as well as defining which sheet to read in from the various country linelists. The second is called cleaning_dictionary.xlsx and contains information for cleaning and recoding variables. Details of the required sheet names and contents are in the table below.

The functions use default to using dictionaries which are pre-defined within the package, however you can also edit these dictionaries by downloading them from GitHub (links provided above), and then specifying the file path within the function call (see help files for function argument details). N.b sheet and variable names should not be changed (otherwise the function will break).

Outputs

2 csv files are outputed from the R functions which merge and clean country linelists (functions called merge_linelist and clean_linelist, respectively).

The first output file beginning with Cleaned_linelist_ (default outputname) contains 43 variables and all cases as merged and cleaned by the R function

The second output begining with ConfirmedCases_ contains 19 variables with only cases that had a confirmed or probable final epi classification. Variable names have been simplifed. This file or files with the same format and variable names as this output should be used in country linelist template (see Templates section).

Descriptions, content, cleaning and other information about the variables in the output csv files are summarised in the tables below:

Table: Required variables by sheet in the linelist dictionary.

Table: Required variables by sheet in the cleaning dictionary

Table: Description of variables in output files.

Templates

There are two templates available for country linelist data. These are similar, however one produces an HTML document (with more content) and the other produces a powerpoint document.

HTML document

The first will produce a HTML output in the style of a report with summary text, tables and figures of the daily situation of Covid-19 in the WHO African region. The dataset beginning with ConfirmedCases_ outputed from the merging and cleaning function should be used as the data input in this template.

Select the template entitled “WHO daily Covid-19 report summary (HTML)”. Change the date of report (dateofreport) and directory and filename of the ConfirmedCases_ file you wish to use (inputdirectory) in the parameters section (params:) at the top in the yaml header and then knit. In order to access templates, follow the steps described in the epichecks readme.

Powerpoint document

This is similar to the HTML document above, but has more tailored content. To access it, select the template entitled “WHO daily Covid-19 report summary (PPT)”.

You will need to update the reference_doc location in the YAML header as well as the current week you are trying to run in the “define_current_week” code chunk. In order to access templates, follow the steps described in the epichecks readme.

Monitoring & Evaluation

Folder structure

Functions and templates are set up to run based on files contained in the COVID-19 Outbreak Documentation > KPI sharepoint folder. This folder contains all the latest M&E files submitted organised in to folders by the month, and sub-folders for the day within that month it was submitted.

File naming and format

File names must start with the three letter ISO code for the country followed by “.KPI.”, afterwhich the rest of the file name can be anything. All files must be in .xlsx format.

For example: A file submitted by Ghana on the 1st of September 2020 would be in the following folder: “COVID-19- Outbreak Documentation/KPI/2020-09/09-01” And would be called: “GHA.KPI. New weekly reporting template.xlsx”

Dictionaries

In order for the functions to work, several data dictionaries are required. The first is called mne_dictionary.xlsx and contains information for defining and naming variables, as well as defining which sheet to read in from the various country linelists. The second is called cleaning_dictionary.xlsx and contains information for cleaning and recoding country names (see country sheet of the cleaning_dict table above). Details of the required sheet names and contents for mne_dict are in the table below.

The functions use default to using dictionaries which are pre-defined within the package, however you can also edit these dictionaries by downloading them from GitHub (links provided above), and then specifying the file path within the function call (see help files for function argument details). N.b sheet and variable names should not be changed (otherwise the function will break).

Table: Required variables by sheet in the cleaning dictionary

Templates

There is one template available for monitoring and evaluation data. This will produce a word document output will summary tables and figures of the indicators for a single country. Figures are coloured using a traffic light system based on the defined cut-offs for the different indicators.

Select the template entitled “WHO weekly M & E Covid-19 report”. Save this in the KPI folder, change the country and week of interest in the “define_current_week” code chunk, and then knit. In order to access templates, follow the steps described in the epichecks readme.

R4IDSR/covidmonitor documentation built on March 29, 2021, 12:05 p.m.