Segmentation Explorer

  • Tool
  • Data Sharing
  • Multidisciplinary Analysis
  • Translation

Better understanding of nuanced differences and similarities between countries, to improve the effectiveness of recommended global policy.

Policy interventions can have widely varying results when implemented in different places. It is challenging to know whether the lack of effectiveness is due to an ineffective policy or an effective policy injected into the wrong environment. Similarly, when a country experiences an effective policy intervention, it may be difficult to know which other countries are the best candidates to replicate that success. This tool provides a data-driven way to discover country-country peer groups – a new way to address these challenges.


This tool is intended to be used by scientists and researchers to explore state, country, or region-level similarities and differences, and help develop hypotheses about policy intervention efficacy.

Access Tool

Demos primary use case and functionality of tool


October, 2018


In Beta Testing


Dave King, Joshua Southerland, Kent Morgan, Jack Kinsey, Derek Grape

The challenge

The challenge that led to the creation of this tool was to determine whether a precision medicine approach could be applied to global policy, resulting in more precise and effective intervention recommendations.

In medicine, outwardly visible presenting symptoms may be driven by variations in underlying biology. Patients who have type 1 or type 2 diabetes may have uncontrolled blood sugar, and the presentation may be similar for HER2+ and HER2- breast cancer. Yet, the optimal treatment for subclasses of these diseases may vary. Before precision medicine, we did not have the tools to tease out the subcohorts that comprised many patient populations; as a result, we provided only partially effective treatments to groups that we wrongly viewed as homogenous.

In healthy birth, growth, and development, we still struggle with an imprecise view of the world that overly homogenizes countries based on outwardly visible factors such as geographic location. However, countries that have similar levels of stunting may vary in other dimensions. A policy intervention that was successful in a country with good infrastructure might not be successful in a country ravaged by war.

It may not always be clear what factors contribute to the environment in which an intervention succeeds or fails. Therefore, we worked to create a new set of tools that would illuminate country-level heterogeneity and lay a foundation for precision policy.

Precision medicine was impossible before advances in genetic sequencing enabled us to see heterogeneity that previously was hidden. Similarly, the first step in achieving precision policy may be the development and use of tools, such as the Country Segmentation tool, to improve clarity about the nuances of regional differences.

Key Features

Use curated datasets or provide custom data

Preassembled curated data sets enable immediate investigation into country-level similarities or differences. The ability to upload your own data sets as spreadsheet workbooks makes it easy to perform custom analysis.

Interactive visualization enables user-driven exploration

Users can explore the country clusters generated by the tool to gain an understanding of the metrics that are the key drivers of similarities and differences.

Focus on longitudinal trajectories at state, country, or region-level

A unique feature of this tool is the clustering based on longitudinal data, not single numeric metrics. Users can explore the metrics that drive the clusters and the individual trajectories for those metrics.

Include geographical distance in segmentation

Users can begin with highly weighting neighboring localities and gradually decrease the importance until only time series data is considered. This helps new users interpret time series similarities for regional events.

Segmentation-metric correlation analysis

Users can explore how the segmentation that results from the selected metric may result in a distinct separation for new metrics that weren’t used as inputs during clustering.

Intervention Opportunities

This tool enables data-driven hypothesis generation about where particular interventions may be most effective, and provides a data-driven view about the countries that may benefit the most from lessons that have been learned elsewhere.

Relational Awareness

The data used to characterize countries and regions are highly multidimensional and longitudinal. This tool provides a way to visualize the meaningful aggregation of hundreds of metrics to provide high-level awareness of key drivers of differences and similarities that may not be immediately obvious by looking at the data in other ways.


Last Updated

September, 2020