Data Upload

This is where the workflow begins. Upload the dataset saved in a comma-separated-value format (.csv); this is the only accepted format. Click on the 'Browse' button to locate the file to be uploaded from your computer; it is as simple as that. Once uploaded, the first 20 rows will be priviewed on the right side.

• Data requirements

  The data to be uploaded must have been tidy up in accordance to the below-described app's input file and input variable requirements:

  • Input file requirements
    • File naming: the file name must use underscore ( _ ) to separate words. Hyphen ( - ) or simple spaces will lead to errors along the uploading process. Consider the following naming example: my_file_to_upload.csv
  
  
  • Input variable requirements
    • Age: values must be in months. The variable name must be written in lowercase ('age').
    • Sex: values must be given in 'm' for boys and 'f' for girls.
    • MUAC: values must be in millimetres. Ensure there are no strange numbers, such as '130.1'. The presence of decimal places will raise error in the data wrangling tab and hault the app.
    • Oedema: values must be given in 'y' for yes, and 'n' for no.

Data Wrangling

Wrangle the dataset for downstream workflow. For this, different wrangling methods are given. Upon completion, this tab's output becomes available in subsequent tabs; therefore, the wrangling method selected herein should match the intended analysis.

Under the hood, the wrangling process consists in calculating age in months and excluding all records that fall under six months and over 59.99 months. Then, it computes z-scores - if either WFHZ or MFAZ method is selected - then it detects outliers based on the SMART flagging criteria - for z-scores. For MUAC, when age in months is not available, values under 100 and over 200 millimetres are considered as outliers. At the the end, two new columns get added into the dataset: wfhz and flag_wfhz for weight-for-heigh z-scores and flagged records, respectively. Moreover, when working with MUAC and when age is available, the following columns get added: mfaz and flag_mfaz for MUAC-for-age z-scores and flagged records, respectively. Finally, when age is not available, only one column gets added: flag_muac which indicates the flagged records based on the above-mentioned criterion.

Once the wrangling process is completed, a preview of the output is displayed on the right side of the tab, wherein the first 20 rows are shown. You can get a full view of the entire dataset by downloading it. Click on the 'Download Wrangle Data' button, and thereafter look for the file in the 'downloads' folder on your computer.

Plausibility Check

As above-described, this tab depends on the previous tab. Select the same method as in the data wrangling. Thereafter, supply the input fields with the corresponding variables from the dataset. For this, a dropdown list of the variable names found in the dataset is given; Select the one that applies, and click on 'Check Plausibility' button thereafter. The app lets you get the plausibility checks results grouped by different categories - up to a maximum of three. For this, supply the grouping variables to Area 1 Area 2 and Area 3

For example: In your dataset, there are the following indentifying variables: 'state', 'county', and 'team'. You wish to get the plausibility check results grouped by province, then by county and then you also wish to check the results by survey teams that worked in each county and provinces. You can achieve this simply by supplying 'province' to Area 1 'county' to Area 2 and 'team' to Area 3 .

Upon completion, you can download the results into Excel by clicking on the 'Download Results' button found at the bottom-right side of the tab.

Prevalence Analysis

As afore-mentioned, this tab depends on the data wrangling. The app lets you estimate prevalence derived from survey and screening. Therefore, the first step is to select the data source - whether survey or screening. If screening, indicate whether there is an age variable in the dataset with values given in months. This tells the app whether to use it or resort to a fallback - when 'No' is selected. In the latter case, you must have a variable called age_cat . This should have the following categories: '6-23' for each record wherein age is between 6 and 23 months, and '24-59' for when age is between 24 and 59 monts. This must be done before uploading the data. The age_cat variable is supplied to the Age categories (6-23 and 24-59) input field. This ensures that MUAC-based prevalence gets age-weighted whenever there is excess of children in the 6-23' category. Read more here and here .

Thereafter, the next step is to select the method to define acute malnutrition. Then, supply the input variables as required.

You can also group the analysis by different groups, as explained in the plausibility check section. Follow the same guide provided therein.

IPC Check

This lets you check whether the IPC Acute Malnutrition evidence requirements for outcome data have been met or not. It checks against survey, screening and sentinel sites data source-related requirements, in accordance with the IPC protocols.

This tab depends on the 'Upload Data' tab; this means that you can check the requirements even before the wrangling the data.

Authorship

This app was developed and is maintained by Tomás Zaba.

License

This app is licensed under the GPL (>=3) license.