Data size difference
When uploading data to Ellipsis Drive the original size of the data is likely to change.
Here we explain:
1. why this happens
2. how you can calculate what size of your data would be in the drive
3. and what you can do to control the storage space you require
Why does my raster data change size?
Data is stored in different file formats that have different compressions. If you use highly compressed data, uploading them to ellipsis drive is likely inflate their size. Visualizations in the drive are PNG compressed. Data itself is lzw compressed. You can choose to remove the data and only keep the visualizations. In many cases this will save up to 95% of storage (read more about this below).
Tile pyramid construction
When raster data in uploaded to the drive, it constructs a tile pyramid of your data. This tile pyramid ensures high performance when scrolling though your map as it serves the right tiles at the right location and zoom level. This does mean that we are storing your data at different resolutions at the same time. As such this may inflate the size of the data you are uploading.
The storage of your project is calculated by the Mercator project surface. It is therefore in your favor to not include unnecessary transparent pixels in your files. For example when uploading data in a file that is largely transparent with just a tiny bit of data in the middle. Your storage will be inflated significantly. Transparent pixels have the same weight as non transparent ones.
The drive re-projects your data to Web Mercator. This reprojection can change the size of your data as it effectively changes the surface area that your data covers. Closer to the poles means covering more surface (making you data bigger), while moving towards the equator mean covering less (making your data smaller).
Why does my vector data change size?
Ellipsis Drive make your vector data easily and flexibly queryable with high performance. To make this possible it stores references to your vectors in multiple indexes. These references themselves have sizes. In case of points the relative size of these references is higher then for complex polygons.
We store additional information about a vector. For example the date it was added and the user who did so. This leads to a small increase in size as well. Again, for points this relative increase is larger then for complex polygons.
How can I know what the size of my data would be in the drive?
Since files are not stored as the files themselves but transformed to a more abstract web service it is not possible to fully predict the exact size the project will take.
The best thing you can do is taking a sample from your data, upload it to an empty project in the drive and check the size of that project in your storage dashboard. Now you know with what factor data size changes when uploaded to the drive. If you upload similar files with similar compression you can expect similar results.
How can I control the storage space required?
The best ways to save storage for raster data are:
1. Remove raw data. This will remove the lzw compressed tile pyramid and can clear up to 95% of your data. This is only an option if you are interested in the visualizations only. If you also want to query the actual raster values you cannot remove the raw data.
2. Make sure your files do not contain large amounts of transparent values.
3. Permanently remove captures that you have deleted. If you delete a capture it is only archived, in the Danger Zone (in the General tab of Map settings) you can permanently delete one.
The best ways to save storage for vector data are:
1. In the Danger Zone you can permanently remove all deleted vectors. When you delete a vector it will only be archived and it’s storage will remain.
2. Do not activate filters on properties. This saves the number of indices in which we need to store references.