# "Communication Style" as part of a user's profile

**URL:** <https://meta.discourse.org/t/communication-style-as-part-of-a-users-profile/285070>\
**Category:** Feature\
**Created:** [November 10, 2023, 1:31pm UTC](https://meta.discourse.org/t/communication-style-as-part-of-a-users-profile/285070 "2023-11-10T13:31:41Z")\
**Posts on this page:** 1\
**Showing post:** 21

<div class="post-metadata">

**Author:** ![sam](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/sam/32/102149_2.png) [@sam](https://meta.discourse.org/u/sam)\
**Post date:** [January 3, 2024, 11:03pm UTC](https://meta.discourse.org/t/communication-style-as-part-of-a-users-profile/285070/21 "2024-01-03T23:03:23Z")

</div>

> [@Saif](#):
>
> We currently experimenting with sentiment analysis on user profiles (only staff can see it).

Note, given:

1. The feature was somewhat unfinished - it supported no drill down so the labels felt very random
2. The feature leaked into the public by mistake due to an internal bug
3. We are not even sure the feature provides value

We rolled back this change for now and are regrouping.

Overall I like sentiment analysis, but its a very tricky feature to get right, some places that it can potentially help moderation:

- Sentiment on the forum this week is highly negative - cause a ton of people are complaining about topic X and Y.
- User X who is usually highly positive is very very negative this month - maybe reach out?

Etc…

Less interested in labeling a lot more interested in “anomaly” detection or overall bad trends.

---

_[View the full topic](https://meta.discourse.org/t/communication-style-as-part-of-a-users-profile/285070)._