Please first proceed to country specification panel to specify the meta data for the analysis, including:

• Specify Country/Survey : Select a country first; then, surveys available for that country will be displayed.
• Choose an Indicator : Indicators are organized by the chapters they appear within the DHS final report. For a detailed list of supported indicators, visit the Tool Kit tab.
• Select Administration Levels : Choose the administrative levels (e.g., Admin-1, Admin-2) for your analysis.
• Display Boundaries and Check Region Count : Use our mapping tool to ensure the administrative regions match your analytical needs. Default boundary sources are GADM or WHO, depending on the app version. For custom boundaries, please contact us.

Access the Data Upload Panel after specifying the meta data for analysis.

• Data Upload : Click Load Data , the app will automatically load the analysis data of selected indicator. Once the data loading is complete (indicated by the green check marks on the upper right of the tab), the app will automatically prepare the dataset for analysis.
• Cluster Map : Use this visualization tool to examine cluster density across Admin areas. Hover to see number of clusters and evaluate data sparsity issue.
• Data Preview : The user may preview and download the prepared analysis dataset in .csv format.

Before proceeding to model fitting, we strongly recommend familiarizing yourself with the methods used in our analysis.

Follow the steps below to carry out analysis:

1. Select from various model types across different administrative levels. Recommended models are listed for guidance.
2. Data Sparsity Check : to assess data sufficiency for model fitting. Results are categorized as:
   • Green check message : Data sufficient. Model is ready to be implemented.
   • Orange warning message : Potential data sparsity issue. Fitting the model is possible but advised against.
   • Red warning message : Data sparsity check failed. Model fitting is not recommended and may result in errors.
   Note: Admin-2 level analysis typically lack sufficient data, requiring Small Area Estimation (SAE). Contact us if you encounter non-meaningful error messages.
3. Following the data sparsity check, choose to
   • Run Models that Passed Check (conservative and recommended option) or
   • Run All Selected Models (extra caution needed for results interpretation).
4. Model fitting statuses are:
   • Successful : The model is fitted successfully.
   • Model not fitted : The model was not chosen for fitting.
   • Model will not be fitted : The model fails the data sparsity check and will not be fitted.

Before using the visualization tools, we strongly recommend reviewing our comprehensive visualization interpretation guide. Access the tools under the results visualization tab to explore subnational mapping and other analytical plots.

• Single Map : Map subnational estimates with four key statistics:.
  • Mean : Point estimates across regions.
  • Coefficient of Variance : Measure of relative variability among estimates.
  • Width of 95% Credible Interval : Reflects the uncertainty around the estimates.
  • Exceedance Probability : Probability that prevalence exceeds a certain threshold.
• Map Comparison : Visualize different statistics from the same model or compare the same statistics across different models.
• Scatter Plot : Compare estimates from different models for the same admin level, and evaluate model performance.
• Ridge Plot : Compare full distribution of estimates across regions.

Tabulated results can be exported as .csv through result tabulation.