Dashboard
Today's open-source intelligence picture
Ingestion volume — last 14 days
High relevance — discovered in last 48h
Items NIGRANI added in the last 48h — the item's own date/time may be older (common for X, whose search can surface a post up to 7 days after it was made).
Map
Geo-tagged activity across Punjab, Rajasthan and Gujarat
Browse & search
Search the full intelligence record by keyword, tag, region, or date
Pattern analysis
Aggregate trends across regions and factions — no individual-level tracking
Active alerts
Explore & trends
Notable sentiment
Notable severity
Notable accounts
Regional & faction activity
Active alerts
Statistical spikes in geo/faction/keyword mentions compared to their normal baseline — a different detection method from the notable sentiment/severity tabs, which flag individual extreme items rather than an unusual volume pattern.
Explore by state or faction
Not tied to an alert — pick any state and/or faction directly to plot or report on everything currently tagged with it.
Sentiment & severity trend
Notable public sentiment
Individual public posts/articles showing extreme sentiment toward the armed forces, with signs of genuine reach — not a profile of any person or account. Deliberately a high bar: most items, even negative ones, won't appear here.
Notable high-severity events
Individual reports of ambushes, attacks, or casualties with signs of genuine reach — independent of sentiment toward the armed forces.
Notable accounts
X/Telegram/YouTube ACCOUNTS (not individuals — a handle is not a confirmed identity) with a repeated pattern of hostile-sentiment or high-severity posts, or manually added by an analyst for surveillance from any item's detail view. Organized by platform below.
Regional activity (7 days)
Faction activity (30 days)
Faction–region co-occurrence (30 days)
Which factions are mentioned together with which states — pairs with at least 2 co-mentions, highest first.
| Faction | State | Co-mentions |
|---|
Predictive analysis
Forward-looking risk indicators, decision support only — see Safeguards for what this is not
Overview
Area historical profile
Hotspot forecast
Faction watch list
Narrative heat
Account behaviour
Cross-sector
Risk calendar
Coverage gap
A risk score here is a probability estimate built from patterns in open-source reporting. It is decision support, not a forecast of certainty, and does not by itself authorize any operational action — see Section 7 of the design note for the full safeguards.
Top hotspot forecasts (Module 4.2)
Top faction watch list (Module 4.3)
Next risk calendar windows (Module 4.7)
Module 4.1 — foundational layer, not itself a forecast. Each district's long-run insurgency/terrorism/espionage history, sentiment trend toward the armed forces, and severe-event record, recency-decayed so a sustained real cooling-off measurably lowers a district's score over time. Every other module below weighs its live signal against this.
| District | Historical tier | Insurgency/terrorism/espionage | Sentiment trend (avg) | Severe-event index | Total severe events |
|---|
Module 4.2 — live trend-slope (severity, sentiment, volume velocity) blended with Module 4.1's Historical Threat Index as a Bayesian-style prior. The same current trend can land in a different tier depending on that district's history.
| District | Risk tier | Trend | Severity slope | Sentiment slope | Volume velocity | Historical prior |
|---|
Module 4.3 — mention trend-slope, status changes, and a small analyst-curated calendar of faction-specific anniversary/trigger dates, deliberately rule-based rather than a black-box model.
| Faction | Status | Resurgence tier | Recent mentions | Baseline | Next trigger date | Status changed recently |
|---|
Faction trigger dates (anniversaries)
| Faction | Label | Date |
|---|
Module 4.4 — exponential smoothing on daily sentiment per district/topic, judged against that same district/topic's OWN historical baseline (district baseline from Module 4.1; topic baseline computed the same decayed-average way) — never one flat platform-wide threshold.
| Dimension | Value | Heat tier | Smoothed sentiment | Baseline sentiment |
|---|
Module 4.5 — a simple historical-rate calculation (no new model), ranking already-tracked accounts (see Pattern analysis > Notable accounts) by how likely their next post is flagged, purely to prioritise limited analyst review time. Entity-level only — never a judgement about a person.
| Account | Platform | Priority score | Flagged ratio | Total items | Posting cadence/wk |
|---|
Module 4.6 — unique to NIGRANI's federated five-sector architecture. Time-lag correlation between this sector's and a peer sector's recent faction/keyword mention activity. Requires SYNC_SHARED_KEY and at least one active peer server (Admin > Network & sync).
Module 4.7 — rolling 90-day forward risk calendar combining generic trigger dates below and Module 4.3's per-faction anniversaries, each with a historical correlation count drawn directly from Module 4.1's severe-event chronology.
| Date | Days until | Label | Scope | Historical correlation |
|---|
Generic trigger dates (elections, festivals, AFSPA review windows, ...)
| Label | Category | Date | Scope |
|---|
Module 4.8 — several sources tied to the same region failing simultaneously, a different pattern from any one source's own health warning. A sudden information vacuum is itself a signal.
Social search
Live search against X — results cover the last 7 days only, and nothing is saved until you import it
Team chat
LAN-wide discussion for all NIGRANI users
Daily reports
Auto-generated each day at 06:00 IST
Commander's Briefing
A one-page summary for circulating up the chain — headline stats, trend direction, and top events. Not a substitute for the full daily report below.
Admin
User accounts, taxonomy, and audit log
Account
Change your password
Password and two-factor authentication settings live in one place for every user now, rather than duplicated here.