-
Notifications
You must be signed in to change notification settings - Fork 1.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Basic taxonomy on events & properties #4267
Comments
Worked on some mockups to work out the UX here. Looking for feedback. One point in particular that's still causing me some issues is on what to do about properties. The problem is that most of the time properties will only make sense to be defined in the scope of an event (i.e. just duplicating events behavior for properties wouldn't make much sense). This however requires implementing full taxonomy on events too. Any thoughts? Events List
Event DetailsEvent PickerAction Picker
|
Re taxonomy, one product which does this is https://www.amundsen.io/ - worth taking a look, perhaps there's worthwhile ideas to share from it. :) Personally things like owner, frequent users, property schema per event with examples (!), date range all seem like worthwhile things. |
I love the events mockups. We actually find that we have a lot of overlaps for properties across multiple events. So for us it would make sense to just have a flat hierarchy of event properties and then maybe even let the "link" between events and event properties be made by posthog itself based on what it sees on a sample of recent events. For example i will probably just go in and add descriptions to the top N event properties that really need them. And then posthog could tell me what events these event properties are found in. So i don't even have to maintain the relationships somewhere explicitly (which will be overhead for me) - posthog will just see what properties are in what events and surface that to me in some way if deemed useful. |
one last idea - being able to publish or share this publicly could be useful too. So some way to either publish some events or properties such that they could be shared with users or the public to say - "look this is all the events and what's in them that we capture about you when you use our system". Might be a bit of a distraction for now but was thinking being able to use all this in some way as part of your own documentation and disclosure could be useful for some too. |
Thanks for the thoughtful feedback @andrewm4894 & @macobo, I've incorporated your input and updated the mockups (latest version on Figma). CC @liyiy Event Picker & Action Picker
Events List
Event Details
Properties
|
looks great. for the properties within the event it could be useful to be able to see for what % of sampled events each property was present on. just might help you quickly debug or spot if you see that say only 5% of the events have a specific property then that might be something i'd need to check on for example. not sure if this might be expensive to compute or not but think it could be handy to have in the popup table for the event properties. |
Thanks @andrewm4894! Makes sense, that sounds indeed quite helpful. I'll leave this to @liyiy to see if it's feasible to be done in the first iteration of this, or if we need to consider it as an improvement. |
Did not mean to close this issue by linking it to that PR lol, reopening it |
@paolodamico I took a stab at exposing some of the taxonomy in an updated concept for picking actions and events. This doesn't capture the full detail you have, but I think it could be enough to make each option more identifiable, while still functioning within the context of other workflows (authoring funnel steps for example). It takes the current picker and organizes options into category tabs. This would enable a user to use a keyword search within a single category. In this concept I am trying to balance describing options in sufficient detail for the current context, while allowing a user to view more event/action details in a secondary context. This all happens in the context of picking an option, so we still need to optimize for that task as well. Thoughts? |
@clarkus I love this! It brings together the best aspects of the old/more compact dropdown and the new, I like that we might re-introduce the tabbed layout, because if a user has a lot of properties, the above design makes the user scroll to see all the categories. |
❤️ it! Some thoughts:
Not sure if you want me to comment on the UI just yet? |
Yes! I was thinking about the work you mentioned on infinite scrolling property lists and such. This would lend well to that loading behavior while still letting a user quickly see that some sets are entirely empty. |
I think we can extend the summary cards shown on hover to include descriptions, tags, etc. I am tracking this today and should have more to share before my day ends.
Thank you for the context and clarification. I will correct this.
I think this can be in the summary card shown on hover. I'm concerned that an action that leads to a separate summary will compete with the actual selection of the item for the current task. I think the designs above are right on target, but I'd like to update them a bit to be more consistent across composition areas.
Thank you. I will add this today.
This is taken directly from the rolling metrics you have in the work above. I agree the labeling could be a bit clearer, but I'm also questioning placement. I will try a few more ideas on this.
+1
I am not sure. I think the use of our branding here is a bit problematic, but I don't have a better solution in mind yet. I will spend some time on icons today.
@mariusandra had similar feedback at #2760 (comment). This bears more research. I think arrow keys are generally a safe bet, but I want to educate myself a bit more and see if anyone else has figured this out. :)
Yup I think that's a reasonable change to make. The label is also much longer so kind of out-competes the other tabs.
Everything is fair game for comments. Feel free to jump into figma and leave comments there if you want to annotate specific items. |
Here's a quick progress update. I've worked through a few icon iterations to help show type families in the combined list of options. I've also included a card shown on hover that will display the object details and provide a secondary action to view all those details directly. This is all hanging off of a filter mechanism, but the list functionality should be the same across instances. |
Really love this new UI! Will add some UI-related comments on Figma, but looking pretty solid 💪 |
Started building something similar here last week: #4844 (take a look at the video if you can). I am a fan of keeping the textbox looking as much like a textbox as possible. Users with hundreds / thousands of property definitions should immediately recognize that they can type to search. 🔍 |
I believe this can now be closed with your work on taxonomy and Marius/Sam's on the event/property selector right @liyiy ? |
Is your feature request related to a problem?
If I work in a larger team or have a ton of events/properties, it can be hard to keep track of what each event and property is. It's only natural I'd use PostHog as the source of truth for this.
Describe the solution you'd like
I'd like to be able to add descriptions (and potentially tag) events & properties to easily identify them. We need to figure out first if this is the best way to support taxonomy (from experiment with dashboards).
This is just a concept, we need to dive a little bit more before building this.
Additional context
cus_IViwMhqr5dnHPa
).Thank you for your feature request – we love each and every one!
The text was updated successfully, but these errors were encountered: