Datasets:
Contributing Parquet dataset
Hi,
Thank you for the great dataset! We have been using it as part of the Urban-M4 project, which is using street view imagery to extract building material properties. As part of this project, we created a Parquet dataset, which combines several of the CSVs together from global streetscapes, to reduce size and make it a little easier to subset. This code to create the dataset can be found here. Would you be open to us contributing this dataset to your project?
Hi Claire,
Please let us know when you decide to publish this dataset. We at the Planbureau voor de Leefomgeving are actively looking for this type of dataset as well.
Best wishes,
Irena Itova
Hi Claire,
Thank you for your interest to contribute! A parquet dataset is a great idea. Also, your Urban-M4 project looks really cool and we’re glad to see our dataset contributing to it.
We have also used this opportunity to discuss among ourselves about a contribution guide, which we just updated on the Dataset Card. Please refer to the guide for the overall steps. For the specific changes, please create a folder named ‘parquet’ under the folder ‘data’ and place the parquet file in it:
Global Streetscapes
├── data/
│ ├── 21 CSV files
│ ├── parquet/ (New)
│ ├── Combined Parquet file (New)
├── …
…
Include in your PR an informative description of your changes (e.g. which several CSV files the parquet file combines) and how you would like to be credited, and we will reflect them on the Changelog on the Dataset Card. Please also ensure that the parquet file contains complete data from the CSV files to be combined to avoid introducing unintended missing data.
Feel free to let us know any further questions. Looking forward to your contribution!