Software

HMigD I App: the Human Migration Database I

Abstract

The HMigD I App (Shiny application https://maciej-jan-danko.shinyapps.io/HMigD_Shiny_App_I/) presents the assumptions and results of Bayesian models used to estimate migration flows. These models integrate administrative data and Labor Force Survey (LFS) data while addressing data quality issues and discrepancies in migration definitions. The application provides reliable and harmonized estimates of bilateral migration flows for 31 countries, covering the period from 2002 to the most recent available data.

The HMigD I App has been developed with support from the Horizon Europe program via the PREMIUM_EU project (Policy REcommendations to Maximise the beneficial Impact of Unexplored Mobilities in and beyond the European Union).

Overview of Features and Capabilities

• Data sources and quality assessment: The app provides an overview of the input database (administrative data and LFS), including the results of data quality assessment used in the model. The data quality issues include:

  1. Accuracy issues: Random errors in data collection affecting population and migration registers and survey data.
  2. Undercounting: Non-systematic bias in migration estimates due to registration failures and non-response in surveys.
  3. Inconsistencies in coverage: Systematic biases resulting from rules in the data collection process that exclude certain population segments.
  4. Inconsistent migration definitions: Varying international duration of stay criteria.
      

• Covariates: The app provides an overview of the different gravity covariates used in some models.

• Results inspection and user survey: The app allows for inspecting estimated flows of different models by comparing them to raw input data and their quality features. Additionally, the app provides a feature where users can share their thoughts on specific flows.

• Model comparison: The app allows for comparing results obtained from different models to assess the effectiveness of different modeling approaches.

• Flow aggregation: The app enables the aggregation of migration flows from different countries using various scenarios related to the freedom of movement covariate.

• Circular flows plots: The app provides circular flow plots for selected countries and years.

• Model mixing: The app allows for combining the results obtained from different models.

The app also serves as supplementary material for the paper titled "Modeling international migration flows by integrating multiple data sources" [in preparation].