6  Survey data

This chapter is under review and may need revisions.

Notes:

6.1 Why does survey data needs its own chapter?

Survey data presents unique challenges and opportunities when working with MOSAIKS. Unlike many other data sources that may be consistently gathered through automated systems or administrative records, survey data:

  • Captures detailed household and individual-level information that’s otherwise unobservable
  • Often follows complex sampling designs that need special handling
  • May have inconsistent geographic referencing across different surveys
  • Requires careful consideration of privacy and ethical concerns
  • Can be expensive and time-consuming to collect, making validation of MOSAIKS predictions particularly valuable

6.2 Types of survey data

Several major categories of surveys are commonly used with MOSAIKS:

6.2.1 Household surveys

  • Living Standards Measurement Study (LSMS)
  • Demographic and Health Surveys (DHS)
  • Multiple Indicator Cluster Surveys (MICS)
  • Labor force surveys
  • National census data

6.2.2 Agricultural surveys

  • Agricultural censuses
  • Crop cutting surveys
  • Farm management surveys
  • Agricultural household surveys

6.2.3 Economic surveys

  • Enterprise surveys
  • Market price surveys
  • Consumer expenditure surveys

6.3 Accessing survey data

Survey data access varies by source and type:

6.3.1 Public repositories

  • World Bank Microdata Library
  • IPUMS International
  • DHS Program website
  • FAO statistical databases

6.3.2 National statistical offices

  • Census bureaus
  • Agricultural ministries
  • Economic agencies

6.3.3 Research institutions

  • Universities
  • Think tanks
  • Research organizations

6.4 Working with survey data

6.4.1 LSMS data

The Living Standards Measurement Study (LSMS) requires specific considerations:

  • Complex multi-topic household surveys
  • Detailed geographic information
  • Panel structure in some countries
  • Integration with agricultural data
  • Varying spatial referencing methods

6.4.2 DHS data

The Demographic and Health Surveys (DHS) present unique characteristics:

  • Standardized across countries
  • Cluster-based sampling
  • Displaced GPS coordinates
  • Rich health and demographic indicators
  • Regular update cycle

6.5 Remote sensing informed survey design

MOSAIKS can enhance survey design in several ways:

6.5.1 Pre-survey planning

  • Optimize sampling frame using satellite-derived information
  • Identify areas of interest based on physical characteristics
  • Stratify sampling based on predicted characteristics

6.5.2 During survey implementation

  • Validate location information
  • Guide field teams with up-to-date imagery
  • Monitor survey progress

6.5.3 Post-survey analysis

  • Validate survey responses against satellite indicators
  • Fill data gaps in hard-to-reach areas
  • Create high-resolution predictions from survey samples

Example of using MOSAIKS features for survey planning:

This integration of MOSAIKS with survey data represents a powerful approach for both enhancing traditional survey methods and extending their reach through satellite-based prediction.

Looking forward

In the next chapter, we’ll look at practical guidance for preparing label data for use with MOSAIKS, including data cleaning, aggregation, and joining to satellite features.