You know what’s awful? Emailing giant CSVs around. A giant CSV is a curse. It’s not a deliverable, or a work product – it’s an assignment, a dare, a boss level challenge to your client to see if they can use it.
Recently at Gigasheet, we’ve rolled out a combination of features that, combined, our customers have begun to use for Data Delivery. Turns out delivering easily analyzed datasets is way better than just sending download links to multi-gigabyte CSV files!
The way data delivery works is this:
1.) You start with a giant, raw dataset. Gigasheet thrives on giant data sets! A billion rows, no problem.
2.) You are asked to cull that down to a more discrete, targeted data set. Again, thanks to Gigasheet’s filtering capabilities, you can do that easily.
3.) You deliver that culled data set to your customer. Not a CSV file, but an actual, usable dataset (complete with an intuitive interface for exploratory data analysis). This is where the latest Gigasheet features have been so helpful.
This new use case has come about because of three new features:
Let’s walk through how this works. For this example, we’ll have an (anonymized) master dataset of contact information, and requests from three clients looking for help targeting their customers: a roofing repair company in Tampa, FL, an upscale tennis club in Seattle, WA, and a Solar Panel installer in Phoenix, AZ.
Here’s what the master dataset looks like (you can view it without logging in):
You see 7,165,725 rows of contact information, probably created after joining a couple open-source datasets. This information includes name, location, year of birth, years experience, implied salary, etc.
But our theoretical clients don’t want to pay to contact seven million people. A roofing company in Florida has no need for the contact info of people in Utah. I, a data broker, have no interest in doing the database administration needed to query this dataset via SQL. That’s where our filters come in.
Simple enough to filter down to 83,464 matching rows for the contacts in the Tampa area. So far, pretty standard.
What Gigasheet now allows you to do is Save this filtered result set as a new sheet. It’s as easy as File -> Save As
I can quickly jump over to my library and confirm the file only has the eighty-three thousand rows of interest, while my master dataset is still available for filtering for the next client.
So for the upscale tennis club in Seattle, we build our filter to include the implied salary field, as it’s targeting a group with more discretionary income:
That gets us just over ten thousand rows, which we’ll again save as a deliverable dataset.
File -> Save As again!
For the Phoenix solar panel client, we’ll filter to anyone in Arizona, and use the multiple columns containing “self-reported interests” to people who have expressed an interest in Green Energy or the Environment.
Again, 18,602 matching rows, again saved as its own sheet
One more File-Save As
And now we have three distinct datasets for three separate audiences.
The previous way of dealing with this would be to download the file as a CSV, upload it to Google Drive, mark the CSV as public, and then email the link for the CSV to be downloaded. Every step there is unnecessary friction.
Here’s what we now do: share the file directly with the intended recipient. They do not even need a Gigasheet account! Simply hit Share
and set the General Access to Public. Now anyone with the link can view the file!
I take the respective links for the respective datasets and deliver them to the intended recipients. Here’s what they see, assuming they do not have a Gigasheet account:
Not a link to download the data. Not a summary of the data. The actual data.
From this view, the anonymous user can filter, pivot, group, sort, and otherwise interrogate the data. No more looking for a tool that can handle the large dataset.
At this point, if the recipient wants to download the dataset, it’s simply a matter of creating a Gigasheet account.Whether the recipient wants the entire file or just a portion, it’s a matter of applying further filters, exporting, and downloading.