Say “Hello, again!” to Deleted Datasets
The possibility (and actuality) of accidentally deleting data is a source of panic for many. To ease that stress, we’ve made it easier than ever to re-unite with deleted datasets.
To recover data, users no longer need to contact our support team to get assistance. Instead, Administrators can head over to the “Data Jobs” section in site Administration to easily restore any dataset deleted within the last five days.
For those with a large number of frequent data updates, we also recently added a date filter to easily find all the data updates, or deletions, within a specific timeframe.
Read this Knowledge Base article for more details on how the Data Jobs page works.
Importing Data Just Got a Whole Lot Faster
Despite the rapid increase we’ve seen in automating data updates, many datasets are still imported by users logging into their Socrata platform and using the import/update wizard. During import, processing the data file could take a while, especially for very large datasets.
To reduce the amount of time spent waiting patiently for data to load, and get data publishers quickly moving on to more high-value tasks like data visualization and analysis, we are excited to announce performance improvements to the Socrata import wizard. Importing data through the import wizard is now up to 9 times faster than previous import times!
To provide some examples, we profiled the performance across 10 different files and the improvement is shown below. Keep in mind that import speeds can vary across datasets and is still dependant on a variety of factors besides the shape of the data.
By moving data through the import process quickly, data publishers will be whisked from step to step, leaving little time for coffee breaks or jumping to other tasks. So fuel up before importing data, and happy publishing!
Small Updates & Bug Fixes
- Data Lens table card now preserves comma formatting from the tabular grid view of a dataset
- Exporting a Data Lens card as a PNG no longer includes a scrollbar on the image
- Choropleth maps built atop very large datasets no longer fail to load
- A spatial lens dataset that could not be made private can now do so