| Election type | Ordinal |
|---|---|
| Culture | Real-Life (beyond PrefLib) |
| Candidates | {10} |
| Voters | {5, 7, 10} |
| Original data is much larger (2600 and 4430 votes) but sampling is used to get final profiles | |
| Instances | 80 |
| Somewhat nonstandard process of generating profiles with partial reuse of data. Depending on the setting, we have 50, 80, or 100 profiles, with some correlation (see Sec. 4, “Datasets” in the paper) | |
| Parameters | None |
| Notes | Participatory budgeting data from Boston 2015 and 2016 (4 approval and knapsack voting). The paper is as much about ordinal as approval preferences. |