Facilitating the data analytics of others


Statistical institutes - regardless of whether they focus on providing complete data analytics services - might wish to provide standardised components (data access, analysis tools) - either for internal use or for external users of data. By providing components rather than complete output, an ecosystem of data analysts could be fostered. At the same, while the reputational risks is mitigated by not providing any analyses, the mere act of providing certain analysis tools might be construed as non-objectivity.

At this daWos session, we will discuss whether and how to provide data analysis components to others.

Questions that could be addressed include

  • How should statistical institutes improve the accessibility to their data for analysts?
    • Is it mainly a question of improving APIs?
    • Should metadata improve?
    • Should open data standards be adopted?
  • How could access to microdata for external data analysts be improved?
  • Beyond data, should statistical institutes provide analysis tools? For whome? More advanced users would presumably have their own. And how would the maintenance responsibility (and costs) for tools be handled?
  • (How) should data analytics services be made available to users? What kind of interactive/participative solutions for users? Can computing/testing platforms help further engaging external users with methods and tools?