NIGRANI West

Area historical profile

Insurgency/terrorism/espionage index (0–100)

ONE combined score, not three separate counts. It rises when more of this area's past reporting involved a known militant/insurgent faction, or matched terrorism/espionage-linked keywords (e.g. "infiltration", "sleeper cell"). Older activity counts for less than recent activity.

Severe-event index (0–100)

How much severe activity (individual incidents rated 4 or 5 out of 5 for severity) this area has racked up historically. Higher = a worse track record. Like the index above, this describes the past — it isn't a forecast by itself.

Historical tier

Critical / High / Medium / Low — this area's overall long-term risk level, blending the two indexes above with its long-run sentiment trend into one label.

Hotspot forecast

Severity slope (per day)

Is the average "how bad" rating of recent stories in this area rising or falling, and by how much each day? Positive = getting worse day by day over the last 30 days; negative = calming down.

Sentiment slope (per day)

Same idea, but for tone. Negative = reporting is drifting more hostile each day; positive = tone is improving.

Volume velocity

How many stories came from this area in the last 7 days, compared to its normal rate. 1.0x = normal, 2.0x = double the usual volume, 0.5x = half.

Historical prior (0–100)

This area's long-term Historical Threat Index (same number as its Historical tier score) — blended in alongside this week's live trend, so a district with a rough track record isn't graded purely on the last 30 days alone.

Faction watch list

Baseline (mentions/week)

How often this faction is normally mentioned per week, before the recent window. Compare it against "Recent mentions" to see if it's suddenly getting talked about a lot more (or less) than usual.

Account behaviour

Posting cadence (posts/week)

How many times per week this account typically posts. Useful for spotting a sudden burst of activity above someone's normal pace.

Narrative heat

Smoothed sentiment

The average day-to-day tone (on a −5 hostile to +5 supportive scale) for this district/topic, smoothed so one unusually angry or positive single story doesn't swing the number by itself.

Baseline sentiment

That same district or topic's OWN normal/average tone over its history — its personal "usual." A district that's always a bit heated is compared to its own past, not to a calmer district's bar.

Probability, clusters & actionable insights

Probability-style % (e.g. "escalation probability", "resurgence probability")

A transparent, reproducible squashing of the exact same trend-slope already shown elsewhere on screen, re-expressed on a 0–100 scale instead of a raw score — not a black-box machine-learning output, and always clamped short of 0% or 100% so it never implies false certainty.

Cluster (e.g. "Critical & Escalating")

A simple, explainable grouping by tier + trend-direction — not machine clustering. Items land in the same cluster only because they share the same tier and the same up/down/flat direction, so the boundary is exactly as inspectable as the tier and trend labels it's built from.

Recommended action / actionable insight

A one-line concrete next step (e.g. "Escalate for immediate review") mapped straight from an item's tier and trend direction — the same standing guidance the tier legend at the top of this page already states, just attached to the specific item it applies to instead of left as a general rule.

Flag-risk tier (Account behaviour) / Recurrence tier (Risk calendar)

These two modules never had a Critical/High/Medium/Low classification of their own — these tiers bucket their existing numbers (flagged ratio; historical correlation count) onto the same scale every other tab already uses, so every module now has both a continuous number and a categorical label.

Confidence, calibration & ground truth

Confidence: Low / Medium / High (the small "Conf:" tag next to a tier badge)

Independent of the tier itself — this says how much sample size, source diversity, and AI extraction certainty back THIS specific call. A "Critical" hotspot built from 3 items and 1 source is real, but should be trusted less than a "Critical" built from 40 items across 5 independent sources. This never changes the tier or score, only how much weight to give it.

Calibration / track record (Overview tab)

Of every past High/Critical forecast that's now had at least 14 days to play out, how often did a real severity 4-5/5 event actually follow in that same district/faction/topic? Entirely automatic — the platform checks its own past calls against what actually happened next, no manual grading required. Shows "not enough history yet" until forecasts have had time to age.

Field reports

Manually-submitted, human-corroborated ground truth — kept entirely separate from the OSINT-derived items everywhere else on this page. Use this to record what's actually been confirmed (or contradicted) in the field, with its own credibility label (unverified/plausible/corroborated).

Response assets / resource-tied recommended actions

A small inventory of available QRTs, patrols, drones, and civic liaison teams (Overview tab). When a hotspot is Critical or escalating-High, the recommended action names one of these directly if it's marked available in that district, instead of generic advice.

Narrative "Drivers" line (Narrative heat tab)

Tallies that row's own driving items (the same ones the "Driving items" button opens) by their AI-classified event type into security-force-facing (ambush, IED, arms recovery...) vs. governance/political (protests, AFSPA debate...) vs. other — so a heat score's underlying cause is visible without opening every item.

Adversary capability & intent assessment

Capability tier / score (0–100)

Demonstrated ability to act, not just how often a faction is talked about — built from recency-weighted offensive incidents (ambush/encounter/IED) it was tied to, how many districts it's been active across, and its current operational status, reduced when security forces have recovered arms tied to it. Separate from resurgence: a faction can be quiet-but-capable just as easily as loud-but-hollow.

Intent tier / score (0–100)

What a faction appears to WANT — built from recency-weighted coercive/extortion activity, stated political demands reported as news, and whether sentiment on its own coverage is trending more hostile. High capability with low intent, or the reverse, is itself useful information, so the two are always shown side by side rather than blended into one number.

Assessment narrative

An auto-generated, plain-English one-liner explaining exactly which observed signals drove that faction's capability and intent scores this pass — the same transparency principle as the driver-breakdown feature elsewhere on this page.

Scenario / What-if comparison tool

What this is

Lets an analyst override one or more live inputs (e.g. "what if the severity trend doubled") and see what score/tier the platform's REAL formula would produce — using the exact same calculation the live Hotspot Forecast / Faction Watch List tabs run, not a separate approximation. Nothing here is saved to the database.

Baseline vs. scenario

"Baseline" is the district's or faction's current real computed score. "Scenario" is what that same formula outputs once your overrides are applied on top of it. Compare the two tiers/scores side by side to gauge how much a hypothetical change would actually move the needle.

Important limitation

This is a recompute of the existing formula under hypothetical inputs — it is NOT a prediction of what will actually happen, and does not itself trigger any alert, snapshot, or calibration record.