TikTok Regulatory Intelligence & Stakeholder Analytics

Independent Business Intelligence Project | Power BI · Power Query · DAX

An end-to-end analytics project examining 567K cleaned public Twitter/X posts about TikTok regulation in the United States between 2020 and 2024. The project combines engagement trends, account-level influence, collection-period patterns, and regulatory milestones in a four-page Power BI report.

This project analyzes public discussion about TikTok on Twitter/X. It does not use TikTok platform data or represent TikTok's internal analytics.

Project at a Glance

Scope	Result
Analysis period	2020-2024
Cleaned posts	567K
Unique users	314K
Recorded engagement	9.1M
Average engagement per post	16.06
Peak engagement period	Q1 2024
Peak quarterly engagement	2.4M
Dashboard pages	4

Recorded engagement is calculated as the sum of likes, retweets, and replies available in the dataset.

Project Context

Organizations operating in regulated markets need to understand when public attention changes, which accounts attract disproportionate engagement, and how discussion patterns coincide with policy developments. This project converts a large social-media dataset into a regulatory-intelligence dashboard designed to support monitoring and exploratory analysis.

The analysis addresses four questions:

1. When did discussion volume and engagement peak?
2. Which high-engagement accounts were most visible in the dataset?
3. How concentrated was engagement across accounts and collection periods?
4. Which engagement peaks coincided with major US TikTok regulatory milestones?

My Contribution

I independently completed the analytical workflow:

• Cleaned and transformed the source data in Power Query, producing a final analytical table of 567K posts.
• Built a date model linking the cleaned post-level table with a dedicated date dimension and regulatory-event table.
• Developed DAX measures for post volume, unique users, likes, retweets, replies, total engagement, average engagement, followers, and peak quarterly engagement.
• Separated reusable numeric measures from presentation-specific KPI display measures.
• Designed four Power BI pages covering executive trends, account influence, collection-period exploration, and regulatory-event comparison.
• Translated the observed patterns into monitoring implications while avoiding unsupported causal or sentiment claims.

Dashboard Walkthrough

1. Executive Overview

Summarizes cleaned post volume, unique users, recorded engagement, quarterly trends, high-engagement accounts, and peak collection periods.

2. Stakeholder Influence Analysis

Explores the accounts generating the highest recorded engagement and compares audience reach with engagement performance. In this project, account-level results are used as a proxy for identifying potentially influential participants; they are not a formal classification of every user as a stakeholder.

3. Discussion Drivers and Collection-Period Exploration

Examines how engagement is distributed across high-activity collection dates. The page is treated as exploratory because the underlying Source.Name field represents source files or collection batches rather than verified media or content sources.

The legacy dashboard screenshot retains several "Source" labels. In the documented interpretation, these are treated as collection-period or source-file metadata, not as publishers, platforms, or causal discussion drivers. Source-level KPI cards are therefore not used as headline findings.

4. Regulatory Milestones and Engagement

Compares quarterly engagement trends with selected US TikTok regulatory milestones. Q1 2024 recorded the highest quarterly engagement in the modeled period.

Analytical Model

The Power BI model contains:

• all: original imported data
• TikTok_Clean: cleaned post-level analytical table
• Dim_Date: date dimension supporting month, quarter, and quarter-year analysis
• Regulatory Events: selected policy milestones mapped by date
• Measure: reusable numeric DAX measures
• KPI Display: presentation-specific formatted measures

TikTok_Clean connects to Dim_Date, while the regulatory-event table uses the same date structure. This enables engagement trends and selected policy milestones to be viewed on a consistent time axis.

Representative DAX Measures

Total Posts

Total Tweets =
COUNTROWS(TikTok_Clean)

Unique Users

Unique Users =
DISTINCTCOUNT(TikTok_Clean[username])

Total Recorded Engagement

Total Engagement =
[Total Likes] + [Total Retweets] + [Total Replies]

Average Engagement per Post

Average Engagement per Tweet =
DIVIDE(
    [Total Engagement],
    [Total Tweets]
)

Average Engagement per User

The legacy Power BI measure is named Average Engagement per Actor; its calculation is total engagement divided by distinct usernames:

Average Engagement per Actor =
DIVIDE(
    [Total Actor Engagement],
    [Total Actors]
)

Peak Quarterly Engagement

Peak Engagement =
MAXX(
    VALUES(Dim_Date[Quarter-Year]),
    [Total Engagement]
)

Key Findings

• The cleaned dataset contains 567K posts from 314K unique users, generating 9.1M recorded engagements.
• Average recorded engagement was 16.06 interactions per post.
• Engagement was unevenly distributed over time, with several pronounced peaks rather than a stable level of discussion.
• Q1 2024 was the highest-engagement quarter in the modeled period, reaching approximately 2.4M recorded engagements.
• A relatively small group of accounts generated disproportionate engagement, indicating that visibility was concentrated rather than evenly distributed across users.
• Several engagement peaks coincided with major regulatory periods. This temporal alignment supports monitoring value but does not by itself prove that a policy event caused the observed change.

Regulatory-Intelligence Implications

1. Monitor milestones and engagement together
   Use a shared timeline to identify when changes in public attention coincide with policy developments.

2. Track high-engagement accounts, not follower count alone
   Audience reach and engagement contribution provide different signals of account visibility.

3. Prepare for event-concentrated attention
   Public discussion can change rapidly around high-salience regulatory periods, requiring timely monitoring and communication planning.

4. Validate signals before drawing causal conclusions
   Combine social-media trends with policy documents, news coverage, and additional contextual evidence before attributing changes to a specific event.

Tools and Techniques

• Power BI: interactive reporting, data modeling, KPI cards, trend and account-level analysis
• Power Query: source consolidation, cleaning, transformation, and preparation of the final analytical table
• DAX: COUNTROWS, DISTINCTCOUNT, DIVIDE, MAXX, VALUES, reusable measures, KPI formatting
• Analytical methods: time-series comparison, engagement concentration, account ranking, reach-versus-engagement comparison, milestone mapping

Documentation

• Methodology
• Data Dictionary

Data and Limitations

• The project uses a supplied portfolio dataset of public Twitter/X posts about TikTok regulation; it does not contain TikTok platform analytics.
• The cleaned analytical table contains 567K records after Power Query preparation.
• Engagement includes only likes, retweets, and replies present in the dataset.
• Account names and follower counts reflect the supplied records and may not represent current identities or audience sizes.
• The regulatory-event table contains selected milestones and is not a comprehensive legal chronology.
• The project does not perform validated sentiment, stance, bot, network, or causal analysis.
• Temporal coincidence between engagement peaks and regulatory milestones should be interpreted as association, not proof of causality.
• Collection dates and source-file metadata do not identify verified publishers or original information sources.

Author

Wenqing Fu
Business Intelligence · Data Analysis · Regulatory Intelligence