Skip to main content
zenodoopen

GitTables benchmark - column type detection

<p><strong>Note:&nbsp;the download page of the entire GitTables corpus is here: <a href="https://zenodo.org/record/4943312">https://zenodo.org/record/4943312</a>.</strong></p> <p>This dataset&nbsp;represents a <strong>small&nbsp;subset</strong> of tables from <a href="https://gittables.github.io">GitTables</a>&nbsp;curated for benchmarking column type detection methods.&nbsp;This benchmark evaluates systems that match table columns to semantic types from the <a href="http://wikidata.dbpedia.org/services-resources/ontology">DBpedia</a> and <a href="https://schema.org/docs/schemas.html">Schema.org</a> ontologies. It&nbsp;is featured in&nbsp;the <a href="https://www.cs.ox.ac.uk/isg/challenges/sem-tab/2021/index.html">SemTab 2021 challenge</a>&nbsp;(CTA task).</p> <p>This&nbsp;dataset consists of the following files:</p> <ul> <li>&ldquo;tables.zip&rdquo;: directory with a sample of 1101 tables from&nbsp;GitTables.&nbsp;Filenames correspond to table IDs, the first column (without column name) corresponds to row indices, column names are replaced with &quot;col_0&quot;, &quot;col_1&quot;, etc. which match to the targets and labels (semantic types).</li> <li>&ldquo;&lt;ontology&gt;_targets.csv&rdquo;: target columns per table ID, 1 file per ontology (DBpedia or Schema.org): columns are &ldquo;table_id&rdquo; (ignore the &quot;_&lt;ontology&gt;&quot; suffix) and &ldquo;target_column&rdquo; (i.e. the column that should be annotated).</li> <li>&ldquo;&lt;ontology&gt;_gt.csv&rdquo;:&nbsp;ground truth column annotations&nbsp;per table ID, 1 file per ontology: columns are &ldquo;table_id&rdquo; (ignore the _&lt;ontology&gt; suffix), &ldquo;target_column&rdquo;, &ldquo;annotation_id&rdquo;, &ldquo;annotation_label&rdquo;.</li> <li>&ldquo;&lt;ontology&gt;_labels.csv&rdquo;:&nbsp;unique labels present in the annotated tables,&nbsp;1 file per ontology: columns are &ldquo;annotation_id&rdquo; and &ldquo;annotation_label&rdquo;.</li> </ul> <p>The labels (semantic types) from each ontology come from:</p> <ul> <li>DBpedia: only properties from the <a href="http://wikidata.dbpedia.org/services-resources/ontology">DBpedia ontology</a>.</li> <li>Schema.org: properties and types from <a href="https://schema.org/docs/schemas.html">Schema.org</a>.</li> </ul> <p>For the entire&nbsp;GitTables corpus, please refer to <a href="https://zenodo.org/record/4943312#.YZQUWr3ML6Y.">this dataset</a>.&nbsp;Visit&nbsp;<a href="https://gittables.github.io">https://gittables.github.io</a>&nbsp;for more background and contact details.</p>

ShareScore

36/100

Overall dataset sharing score

Score breakdown

These five areas show where the dataset supports — or may limit — practical reuse.

Stewardship
4
Harmonization
4
Access
20
Reuse readiness
8
Engagement
0