Most interesting topics for discussion?

  • 12 September 2019
  • 11 replies
  • 292 views

Userlevel 7
Badge +1
  • Good Looker
  • 1167 replies

Hello clever community members,


What’s the number 1 topic/area of discussion around Looker or data in general that you’d be most interested in when it comes to hearing best practices from experts, sharing knowledge with peers, or getting help?


Ex: “Making things dynamic with Looker”, “Caching for optimal performance”, “Vanity metrics”, and so on.


Truly no topic too big or too small— Just share what you’re most interested in collaborating with others on.


Just asking for a friend 😉. This may turn into something later, but right now, just collecting topics.


11 replies

Userlevel 4

Hello Izzy,


I think it’s a great topic and I’ll start with my 2 cents (I don’t know if it has its place here but I try).


I started deploying Looker some months ago and I have a recurring complaint from users about the lack of interactivity from dashboards, typically:



  • The rigidity of the filter bar on top

  • The lack of fancy controllers/filter dropdown/radio buttons that you can place next to your tiles for example.

  • The lack of clickable filters (things like you click on a value and it filters the complete dashboard)


Most of my users are used to tools like Google Data Studio, Tableau or Qlik and it’s a shame to be limited by some front end missing features when you have the in database for real time, the modelling layer which gives incredible power to develop very quickly some complex dimensions and measures etc…


At the moment, we were already able to bring new stuff and bring new analysis pattern with Looker but we are missing the small cherry on the cake to make users feel comfortable 😉

Userlevel 7
Badge +1

Gotcha, so as a potential subject for a discussion session, maybe something like “Making dashboards feel interactive”? Thanks for your input!

Userlevel 4

Yeah! Exactly 😉

Userlevel 5
Badge

I’d love to see more front-end/back-end integration in Explore environments. For example, when a business user asks “can I get more significant digits there?”, or “I think this would be better if X was called Y instead”, it’d be great to be able to enter a config mode on the Field Picker, change the value formats or labels or group labels, and have that reflect in the LookML. This would bring the opportunity for business users to help bridge the gap with developers who may not be aware of the common business user vocabulary. I also think an embedded data dictionary feature would be huge.

Userlevel 7
Badge +1

Ben, these are great ideas. I’m gonna steer the conversation a bit though— This isn’t intended to be a feature request thread, rather a way to collect topics for potential discussion groups / knowledge shares between Looker users and experts. Things like how to set up caching policies, build engaging dashboards, or spark adoption of data in your organization.


Apologies for not being clear in my initial post! I’ve edited it for clarity.

Userlevel 2

I’d be interested in learning how people are taking advantage of the custom visualizations feature - best practices, tips for developing your own, etc.


And I agree with Anthony - would love a topic on making dashboards feel interactive

Userlevel 6
Badge

We have done well with adoption and satisfaction but our only/main two complaints are the popular one so far regarding dashboards and interactivity and then the other is complaints of “slowness”. This is of course mostly a problem with the model and database however caching can play a huge part, I wrote a great document to cover off a lot of these areas but it is a complicated subject and unless you are fairly experienced with how Looker works and LookML then you are unlikely to get much sorted.

Userlevel 2

I would be really interested in best practices around using Looker for data science and advanced analytics. I’m thinking the “full loop” of A) getting data from Looker to train a model e.g. using R/Python in Jupyter, 😎 displaying results in Looker again, and maybe even C) using Looker in productionizing the model if that’s something folks have done.


There are resources like JOIN 2017 - Deep Dive - Integrating Looker with R & Python, but more discussion and examples from other customers who have made Looker part of their ML environments would be great.

Userlevel 3

Hey Izzy


I think a good topic is how to properly manage a dev and production environment. As you know there is no dev environment for dashboards (like we have with LookML), so what is the best way to manage front end changes on customer-facing front dashboards, particularly when they coincide with LookML changes?

I kinda know the answer as far as what the best practice is (as I think I’ve asked you and others a million times 😉) but would be good to hear in more detail, and this topic is probably relevant to a lot of users.

I second the topic of Ben of an (embedded) data dictionary. I know we have Markdown features, but AFAIK there is no way to filter/restrict in respect to users access rights/roles/models. Descriptions and labels are often not enough for more complex stuff. How do others manage a data dictionary?

We are always looking for better ways to use looker in an automated way when it comes to code changes, build and deploy to many different environments. Subjects about startup options, environment variables, log file locations, project folders, and automated ways to use/setup looker (API), and even docker topics are part of our interests.

Reply