Dataset research report
Sepal Datasets research report
A reproducible data report with schema notes, generated chart evidence, suggested follow-up questions, and export-ready Helix queries.
Executive Summary
SEPAL: Scalable Feature Learning on Huge Knowledge Graphs for Downstream Machine Learning This dataset contains the Mini YAGO3 knowledge graph and the downstream tasks used in the SEPAL paper (https://arxiv.org/pdf/2507.00965). SEPAL is a knowledge graph embedding method for very large knowledge graphs. It is designed to produce good embeddings for downstream regression and classification tasks. The code for SEPAL can be found at https://github.com/soda-inria/sepal… See the full description on the dataset page: https://huggingface.co/datasets/inria-soda/sepal-datasets.
Research Context
Sepal Datasets: 500 rows by 7 columns. These exploratory charts are generated automatically from the data - open the dataset in Helix to ask your own questions.
Data Profile
Chart Evidence
These views are generated from the dataset profile. Each chart is paired with a Helix query so it can be opened, adjusted, and exported.
Follow-Up Queries
Preview Rows
| # | raw_entitytext | yago3_entitytext | yago4_entitytext | yago4.5_entitytext | wikidata_entitytext | freebase_entitytext | targetfloat |
|---|---|---|---|---|---|---|---|
| 1 | New York, NY | New_York_City | New_York_City | New_York_City | Q60 | m.02_286 | 5.855 |
| 2 | Buffalo, NY | Buffalo,_New_York | Buffalo,_New_York | Buffalo_u002C__New_York | Q40435 | m.019fh | 5.334 |
| 3 | Rochester, NY | Rochester,_New_York | Rochester,_New_York | Rochester_u002C__New_York | Q49218 | m.0y1rf | 5.322 |
| 4 | Yonkers, NY | Yonkers,_New_York | Yonkers,_New_York | Yonkers_u002C__New_York | Q128114 | m.0n6dc | 5.78 |
| 5 | Syracuse, NY | Syracuse,_New_York | Syracuse,_New_York | Syracuse_u002C__New_York | Q128069 | m.071cn | 5.234 |
| 6 | Schenectady, NY | Schenectady,_New_York | Schenectady,_New_York | Schenectady_u002C__New_York | Q331380 | m.0fdpd | 5.436 |
Data Dictionary
- raw_entity text
- yago3_entity text
- yago4_entity text
- yago4.5_entity text
- wikidata_entity text
- freebase_entity text
- target numeric
Method And Limits
- Load the catalog entry and preview rows from the processed dataset file.
- Infer numeric, categorical, time, and location fields from real columns.
- Generate a small set of defensive Plotly chart specifications from that profile.
- Expose each chart idea as a query link so the report can be rerun or exported in Helix.
This report is intentionally reproducible. It uses the local catalog metadata and generated chart specifications rather than claiming external conclusions beyond the dataset.