NTCIR-16 Data Search 2

A shared task for ad-hoc dataset retrieval

NTCIR-16 Data Search 2 is a shared task on ad-hoc retrieval for governmental statistical data. The second round of Data Search addresses the retrieval from a dataset collection published by the Japanese government (e-Stat), and one published by the US government (Data.gov). Furthermore, this round introduces two additional subtasks: QA subtask and UI subtask.


data-search-org@googlegroups.com or @NTCIRDataSearch


Jan 31, 2021 Dataset collections and training queries release
Aug 31, 2021 Sep 30, 2021 Registration due and test queries release
Sep 30, 2021 Oct 31, 2021 Submission due for the IR subtask
Nov 31, 2021 Nov 30, 2021 Evaluation results release for the retrieval subtask
Feb 1, 2022 Submission due for the QA and UI subtasks


Please take a look at What participants must do. Then, go to http://research.nii.ac.jp/ntcir/ntcir-16/howto.html and follow the registration instruction. The registration is necessary for participants to submit their runs.


For each subtask, we provide Japanese and English test collections.

  • IR Subtask

    • In this subtask, given a query and a dataset collection, a system is expected to generate a ranked list of datasets.
  • QA Subtask

    • In this subtask, given a question and a dataset, a system is expected to generate an answer to the question, mainly by extracting a part of the dataset.
  • UI Subtask

    • In this subtask, participants are expected to develop a search system with an effective search interface for dataset search tasks.


  • Makoto P. Kato (University of Tsukuba)
  • Hiroaki Ohshima (University of Hyogo)
  • Ying-Hsang Liu (Australian National University)
  • Hsin-Liang Chen (Missouri University of Science and Technology)