Socialdesk

Problem

Brand and social teams jump between Instagram, news feeds, and spreadsheets to answer basic questions: What shipped? What is trending? Where is sentiment going? What needs attention this week? Without a single ops view, coverage gaps and stale scrapes stay invisible until a stakeholder asks.

Role

Product / frontend engineering on a social operations dashboard: information architecture, monitoring UI, trend surfaces, and alert patterns built for day-to-day use—not a one-off marketing mock.

What I built

• Unified monitoring shell — overview, brand analysis, monitoring, trends, and settings in one workspace instead of tab chaos across tools.
• Time-window controls — 7D / 14D / 30D / 60D / 90D filters so teams can move from weekly pulse to quarter view without leaving the page.
• Source-aware pulse — post counts, source composition, top and recent posts so coverage (and missing connectors) is obvious at a glance.
• Topic and content insight — AI-assisted read of content strategy and conversation themes, with a clear “run analysis” action rather than opaque magic.
• Ops alerts — stale scrape warnings, cross-platform coverage gaps (e.g. Facebook still pending while IG and news are live), and comparative engagement signals so the next focus is actionable.
• Responsive dashboard UI — React + TypeScript layouts that stay usable on laptop and wider ops monitors.

Product shape (high level)

Connectors (IG · news · X · …)
            │
            ▼
     Ingest / normalize
            │
            ▼
   Socialdesk dashboard
   ├─ Pulse & windows
   ├─ Posts (top / recent)
   ├─ Trends & keywords
   ├─ AI topic insight
   └─ Alerts

Live demo

socialdesk.mamenesia.com — example brand monitor with multi-source posts, windowed metrics, topic insight, and operational alerts.

Outcomes

• One place to see coverage, trends, and gaps instead of stitching exports.
• Alerts that name real ops issues (stale data, missing platform, engagement skew)—not vanity “you have notifications.”
• Dashboard patterns reusable for other brand or campaign monitors.

Tradeoffs

• Optimized for operator clarity (pulse, posts, alerts) over a full BI suite; heavy custom reporting can sit on top of clean event streams later.
• Connectors and scrape freshness are first-class UI concerns—the product tells you when data is incomplete rather than pretending every source is green.
• AI topic insight is on-demand so cost and latency stay under control for routine dashboard loads.

What I’d do next

Deeper connector health history, saved alert rules per brand, and exportable weekly digests for stakeholders who never open the dashboard.