Yesterday, we saw with a real life use case how the Formula Assistant will help us to build calculated fields faster with a minimum of knowledge about the Looker Studio functions library.
Today let’s continue with another feature preview.
Don’t look for Gemini Features inside your Looker Studio Pro interface, it is not generally available for the moment.
So , let’s see what we could do with another upcoming feature powered by Gemini AI, inside Looker Studio Pro: The automated generation of annotated reports inside Google Slides
And one time again, let’s try to push the limits…We don’t wish just screenshots and descriptive texts…no no no…gap-fill exercise would be too easy.
We want contextual highlights, and we need also trends analysis with insights to review and validate. Yes, we validate any outcome from Gen AI. Let’s keep in mind, it is a catalyst, not a substitute.
This time, let’s feed Gemini AI with a time-series analysis for correlation 😈
I reused the data materials of a blog post inspired by a quintessential conference.
Basically the goal would be to determine if there was a timing correlation between 2 drive-to-store metrics.
So I put it in a report page and start to play with the slides generator.
The result? Watch the demo video below:
– Key variations with values, % of variation and concerned dates
– But also conclusions and identification of general trends…correlation for this use case 😮 😈➡️👼🏻
“Increases in Push notifications typically correlate with increases in Google My Business Store Views, suggesting a potential link between the two metrics.”
…ok ok, it is not just about fast generated supports, it is also about intelligence.
Slides as a end, but probably often as a way to get insights quickly with slide sandboxes.
Promising feature, indeed!
Let’s share a small suggestion to the amazing team working on Looker Studio.
What if the chart type is used as an additional signal to drive the intent of the analysis, and so the slides annotations:
- Google Maps for distance affinity,
- Scatter chart for correlation,
- Time series for anomaly detection
- …
- and Table chart would mean “No idea Gem, be creative and find something for me” 😛
On the other hand, If I use the same data set for several visualizations, this additional signal would help Gemini to choose automatically the right annotation following the viz of the slide. Just a suggestion.
By the way, thanks a lot and again, well done Looker Studio and Looker BI teams.
I’ve noted other suggestions but it is really good, bringing real operational value.
Thanks to this kind of tests, I’ve got in mind a lot of activities where it would be advantageous to put Gem in the loop. Yes, we are close now, I can call it “Gem”….et j’aime!
