Community Analysis

Community Daily Report

Automatically generate an operations daily report from a single day of community data: T-1 vs T-2 day-over-day change, key focus points, real quotes, sentiment summary, risk and compliance, and action recommendations.

IndustriesGamesSocialContent

Agent · Skills Live

skill · loadedindustry+100

Run this Skill on my data

Metric overview

live

DAU

48.2k

+6.4%

Payers

3.1k

+2.1%

ARPU

$0.72

−1.3%

Analysis readyevidence attached

The Community Daily Report Skill is a community-operations daily-report generation capability in ThinkingAI's industry Skill library. It helps operations and community teams automatically generate a structured operations daily report from a single day of community data, completing the T-1 vs T-2 day-over-day comparison, key-focus-point extraction, real-quote citation, sentiment summary, risk and compliance signals, and action recommendations. This is ThinkingAI's decade of industry experience, packaged into ready-to-use Skills.

Live walkthrough

AgentSave as AgentExport session

community-daily-report

Generate yesterday's community daily report. We just pushed a new version the night before last and I haven't had a chance to look at the data: highlight the day-over-day anomalies and players' own words; it goes out to the group at 10.

Time inference: no date specified, defaulting to T-1 = 07-16, T-2 = 07-15

Pull two-day overview metrics (post volume / comment volume / by channel)

T-1 post volume 3,104, +15% day over day

Compare sentiment distribution and flag anomalies

Negative share 20% → 25%

Aggregate trending topics and extract real quotes

Risk-and-compliance scan: file experience and compliance dimensions separately

0 compliance risks, experience feedback concentrated on 1 item

The daily report is generated. Yesterday's only anomaly signal is an Android login crash. It explains both the +15% post-volume lift and the 5pp rise in negative sentiment, while every other topic is within normal fluctuation. The 07-16 daily report for a sample anime card game (demo) follows.

1. Today's top 5 hot topics

Topic	Heat (posts + comments)	Sentiment	Suggested priority
New-version login crash (Android)	1,286	78% negative	High: must respond today
New banner featured-character power discussion	842	55% positive	Medium: steer toward strategy content
Story-chapter update praise	617	71% positive	Low: amplify with official-account fan edits
Gacha animation stutter feedback	433	52% negative	Medium: track together with the crash issue
Complaints about slow support response	289	63% negative	Medium: pass to the support team to check staffing

← Scroll to see more

2. Daily report at a glance

Volume: T-1 post volume 3,104, +15% day over day, with almost all the lift coming from the login-crash topic (412 related posts)

Sentiment: negative share 20% → 25%, driven mainly by the Android login crash and the gacha animation stutter

Quote: an original Baidu Tieba post, 'After the update it's stuck on the logo screen; reinstalled twice and still can't get in': similar descriptions recur frequently in the comments

Compliance: no spam ads or fraudulent content found, the risk surface is clean; the login crash is filed under the product-experience dimension

Recommendation: hand the crash to product and engineering, and get an official statement out before midday to stop the sentiment from climbing

Heads-up

The login-crash negativity is still climbing: since the version push at 22:00 the night before last, the hourly volume of related feedback still hasn't peaked. This morning is the response window: a short notice saying 'issue identified + fix ETA' will hold sentiment down better than fixing quietly, and after it goes out I can keep watching the negative share come back down.

The daily report has been pushed to the 'Community Ops' group on Feishu and archived to the dashboard.

On your data

That was a simulated run

Leave your work email and we will run a live walkthrough on your real business data.

The problem

More than 75% of community daily reports are still assembled by hand: overview metrics, sentiment distribution, daily summaries, hot posts, comment data, and risky content each have to be checked one by one, averaging 20 to 30 minutes. Daily reports often lack a day-over-day comparison, so single-day data alone cannot tell whether a change is anomalous; user feedback is often summarized by operations rather than quoted verbatim, distorting the information; and product issues and compliance risks are mixed together, hindering directed follow-up.

What it does

T-1 vs T-2 day-over-day comparison: automatically compute changes in volume, sentiment, and channel, flagging anomaly signals

Real-quote citation: key focus points come with users' actual words obtained via MCP, not replaced by operations summaries

Sentiment summary table: list triggers and possible impacts separately for positive, negative, and neutral

Risk and compliance separated: product bugs and UI issues go under the experience dimension, and violations, fraud, and sensitive content go under the compliance dimension

Smart time inference: when no date is specified, automatically take yesterday as T-1 and the day before as T-2

When to use it

Generating a community operations daily report for team sync each day

Comparing today's and yesterday's community data to judge anomalies

Extracting key focus points and real user quotes from community data

Viewing a single day's community sentiment distribution and triggers

Quickly identifying community risk and compliance signals

In the field

Case

A game community · daily-report automation

The team spent 25 minutes daily assembling 6 data sources to write the report. The Skill automatically collected T-1 and T-2 data, found volume up 15% and the negative share up from 20% to 25%, with the key focus point being "new-version login crashes" and real Tieba comments attached. The report went from a 25-minute vague manual write-up to a 5-minute structured flash report.

FAQ

What is the difference between a daily and a weekly report?

A daily report is a single-day flash, focused on the T-1 vs T-2 day-over-day change and 1 or 2 anomaly signals. A weekly report is a 7-day summary, focused on core-event arcs.

What are T-1 and T-2?

T-1 is the target day and T-2 is the baseline day. When no date is specified, yesterday defaults to T-1 and the day before to T-2.

Are the quotes users' actual words?

Yes. Quotes come from the original post and comment text obtained via MCP tools, and when there is no suitable quote it is marked as no clearly representative quote, with nothing fabricated.