A spreadsheet is enough for trend research when you need history, source links, and a weekly chart. The hard part is keeping the queries and columns stable so a change in your process does not look like a change in the topic.
This lightweight social media dashboard keeps the underlying rows visible in Google Sheets.
SocialListeningAPI lets Claude or ChatGPT search TikTok, YouTube, Reddit, X, LinkedIn, and Hacker News through the MCP. It returns normalized records with source URLs. An external automation writes rows and builds charts in Google Sheets.
Step 1: Freeze the social media dashboard method
Choose one topic and 3 to 10 query variations. Create these columns:
- Run date
- Query
- Platform
- Item ID
- Source URL
- Published date
- Author
- Available engagement
- Topic label
- Review status
Use the source URL as the main deduplication key. Do not change the query list during a reporting period without recording the change.
Step 2: Collect rows through the MCP
Create a SocialListeningAPI account, claim the 100 free credits, and connect the MCP to Claude or ChatGPT through OAuth.
Ask Claude or ChatGPT for a row-ready result:
Search TikTok, YouTube, Reddit, X, LinkedIn, and Hacker News for each approved phrase. Return one
row per result with run date, query, platform, item ID, source URL, published date, author, and
available engagement. Do not calculate a trend and do not invent missing values.One phrase across these six sources costs up to 7 credits because LinkedIn costs 2 and each other successful search costs 1.
Step 3: Append and deduplicate externally
Use Apps Script, Zapier, Make, n8n, or another external automation to append the returned rows. Skip a row when the source URL already exists. Keep blank fields blank instead of replacing unavailable engagement with zero.
SocialListeningAPI performs the searches. Google Sheets, the scheduler, deduplication, formulas, charts, and notifications remain external.
Step 4: Build a directional weekly chart
Create a pivot table by week, platform, and query. Chart unique returned URLs and, separately, available engagement. Review a sample of rows before calling a rise or fall a trend.
Search results do not represent every public post. Platform coverage, result limits, and query quality can change the counts.
Conclusion
The expected outcome is a simple history table and a weekly chart where every point can be traced to source rows. Move to a custom database only when the sheet becomes difficult to review or update.
Continue with Detect Emerging Category Terms.
