Post

Giving non-technical coworkers a tool that generates reports from database

I have a RedShift data warehouse filled with dozens of different tables. For some processes, a few joins might be needed for pre-processing.
Lately, I've been messaged by several data visualization platforms looking to streamline reporting.

What platforms have you used for your non-technical coworkers? What's the associated price tag per user-type? How much training did it require for you and your team?

Share
maguay's avatar
13d

My first thought here is Google Data Studio—but it doesn’t natively integrate with Redshift, so that may be a bit more trouble on your setup end than other tools. The good thing with Google Data Studio is that it’s incredibly easy for anyone with any G Suite experience to use. Basically, if you know how to use a spreadsheet and create charts from it in a presentation, you can use Data Studio.

Other great options from Capiche discussions about tracking data in dashboards, KPI tracking, and data visualization include Grow, Graphy, Databox, Metabase, Looker, Amplitude, Chart.io, and more were all recommended. Of those, Grow and Chartio both seem to be fairly accessible to non-technical users and have built-in Redshift integrations, so they might be good to check.

5 points
-Lee-'s avatar
@-Lee- (replying to @maguay )
10d

A vote for Grow. It’s simple enough that non-technical users can navigate and transform the data. They will also do an on boarding session with your users as part of sign up. Users are free they only charge for metrics / dashboards.

From an experience point of view however, I would say your biggest issue with non-tech people is them understanding the tables and data.

2 points
maguay's avatar
@maguay (replying to @-Lee- )
7d

I keep hearing great things about Grow!

Great point on it being a challenge to ensure non-technical people understand the data. Do you have any tips to help there, in onboarding team members to data work? That makes me think tools like Grow would do well to have some way to add internal notes/documentation/onboarding stuff so data teams can help people get started once then hand it off.

2 points
-Lee-'s avatar
@-Lee- (replying to @maguay )
6d

For citizen developers I always prefer doing the joins / transforms before hand. Rename columsn to something that makes sense for everyone and less data is more IMO.

3 points
mitchellreynolds's avatar
@mitchellreynolds (replying to @maguay )
12d

This is spot on. Plus G Suite has a unified experience.
I'm not too familiar with Google Data Studio [GDS]. Would I be able to create views within GDS removing PII? Or, would I create a table removing PII to be uploaded into GDS?

Also, I'll check out these discussions; thank you for the links!

2 points
maguay's avatar
@maguay (replying to @mitchellreynolds )
12d

Especially without a built-in Redshift integration, my guess would be you'd create a table without PII/with the data you want to be used first, and import that into Google Data Studio. That, or you'd need to add a 3rd party integration tool to the mix, and see what filtering options it has.

Google Data Studio is great for marketing teams, as basically a better UI for Google Analytics/Ads/etc—beyond those built-in integrations, might take a bit more work to get the data you want in it, but the end result should be equally easy for your team to use.

2 points
dvddmn's avatar
12d

In the company I was previously working, business people were using SQL server management studio with queries saved in .sql files to create a report.

I was in shock when I saw one of them connect to the production DB and opened up a query window.

3 points
mitchellreynolds's avatar
@mitchellreynolds (replying to @dvddmn )
12d

I tried this at first!

I sent some homework for the basics in SQL (sqlzoo) and created a lesson plan. The lesson plan quickly reviewed the basics then dove into real business questions / scenarios.
The problem: Most didn't do the homework (even after multiple follow ups)... And they were lost because 80% of my lesson plan was toward the interesting questions using intermediate queries.
Lo and behold, here we are!

For those learning from my mistake: Make sure you convince the "biggest voice in the room." Even after a follow up 1-on-1 session, he didn't understand thus doesn't recommend SQL. I gave a bad first impression that it was difficult.

1 point
maguay's avatar
@maguay (replying to @mitchellreynolds )
7d

You're absolutely right on needing to convince someone influential. It took a while for my former marketing team to start using Looker effectively, and it all essentially started when one colleague kept mentioning Looker in every meeting. That was enough to eventually evangelize it across the team.

2 points
qthdh's avatar
9d

You might want to give a try to Metabase.
It has direct integration with Redshift.

You need to be technical to install it, but once it’s installed, it’s a blast for both technical and non technical people.

3 points
jeff_graphy's avatar

I don't have a solution for you yet but what you're describing is what we're working on at Graphy – https://www.graphyapp.com. So far we've been focused on direct integrations with tools but we're starting to explore how to bring database connections into the product and make it easy for non-technical users to get the most from the data, with the least amount of SQL possible.

If you're up for it, I'd love to talk to you more about your use case in specific to better understand it. :)

2 points
ginoarendsz's avatar

Hey @mitchellreynolds,

At a previous employer we had a similar setup, from what I recall they also used Redshift. The only thing I can't tell you were the costs, because I don't know these costs :).

Now the tool that was used internally, for people like me that are non-technical-ish, was Saiku (https://www.meteorite.bi/products/saiku/). A super 2000s looking interface, but that was powerful, fast and always got the job done. Basically all of our dimensions and metrics where in there, allowing us to combine different data cubes to extract the data we were looking for. Every new hire got a training of a few days and had to do an exam before they were given access. Not necessarily because they could break stuff, but merely to tackle the simple questions upfront.

For our more technical employees we had Mode Analytics, where they could write their own SQL queries and obviously do more with the different cubes.

But these two really helped our teammates to get the answers they were looking for without disturbing our BI team too much.

Hope this helps!

2 points
rpatools's avatar
12d

JetAdmin.io or Internal.io for sure!

2 points
barrald's avatar
12d

Are you looking to give the users the ability to do the joins? Or just a finished view on top of the processed data so they can visualize / access?

If the latter, Hex (hex.tech) might be a good fit. Happy to hook you up with early access (email me at barry@).

2 points
phionahq's avatar
10d

You are able to do this with Phiona- users can access the underlying Redshift data and join and/or manipulate the data tables without SQL or Python in our platform. Feel free to reach out at steve@phiona.com and I'd be happy to give you a demo!

We do team-based pricing, so for Redshift access and reasonable data sizes / refresh frequencies, it should be pretty cost efficient compared to some of the larger BI platforms.

2 points
How do you automate follow-up emails?

I've got a funny use-case for a client I'm volunteering on. They're dealing with some legacy folks who work in Excel. Every week they send out an email to a series of people with an Excel sheet at...

The community for power users.