Methodology & Safety Data

Last updated: March 2026

Section 01

Introduction

Juliapp is a community-powered safety mapping platform built on the principle that safety information should be transparent, accessible, and community-driven. Our mission is to empower individuals with real-time awareness of their surroundings so they can make informed decisions about their routes, timing, and travel.

Every data point on the Juliapp map originates from a community member who witnessed or experienced a safety-relevant event. Reports are collected through a structured submission process, enriched with contextual metadata (category, severity, time, location, environment), and displayed on an interactive map in real time.

This page explains exactly how data is collected, processed, scored, and displayed — because we believe the methodology behind safety tools should be as transparent as the data itself.

Section 02

Report Categories

When a user creates a safety report (a "pin"), they select a category that best describes the incident. Each category is designed to capture a distinct type of safety concern.

Category	Description	Examples
Harassment	Verbal, sexual, or persistent unwanted attention directed at an individual	Catcalling, verbal intimidation, sexual harassment, following
Stalking	Repeated or persistent surveillance, following, or threatening behavior	Being followed on foot or by vehicle, repeated appearances, surveillance
Dark Area	Environmental hazard related to poor visibility, infrastructure danger, or unsafe conditions	Broken streetlights, poorly lit passages, construction hazards, obstructed pathways
Aggression	Physical violence, robbery, or direct confrontational threat	Assault, robbery, bag snatching, physical altercation, armed threat
Intoxicated	Presence of intoxicated individuals behaving erratically or aggressively	Aggressive drunk behavior, drug use in public, erratic/unpredictable conduct
Other	Any safety concern that does not fit the above categories	Suspicious activity, vandalism, unusual gatherings, animal hazards

Categories are intentionally broad to remain inclusive without being overwhelming. Users can add a free-text description to provide additional detail beyond the selected category.

Section 03

Severity Levels

Each report includes a user-assessed severity level that communicates the perceived level of danger. Severity affects how prominently a pin is displayed on the map and how it factors into route danger scoring.

Level	Label	Indicator	Description	Examples
1	Low / Mild	Green	Minor concern, no immediate physical danger	Suspicious loitering, minor vandalism, broken streetlight
2	Medium / Moderate	Yellow	Moderate risk, heightened awareness recommended	Verbal harassment, petty theft area, aggressive panhandling
3	High / Danger	Orange	Significant risk, active caution advised	Repeated theft incidents, aggressive confrontation, stalking pattern
4	Critical / Emergency	Red	Immediate danger, emergency-level threat	Assault in progress, active emergency, armed threat

SOS alerts are a special category of critical-severity reports triggered through the emergency button. They broadcast the user's live location to trusted contacts and nearby community members, and are styled with high-visibility pulsing indicators on the map.

Section 04

Environment Context

Reports include an environment context that describes how the reporter was traveling at the time of the incident. This metadata enables time-and-mode-specific safety analysis.

Environment	Description
On foot	Walking or running in any outdoor environment
Transit	Using metro, train, or tram systems
Bus	On a bus or at a bus stop
Cycling	On a bicycle, scooter, or similar micro-mobility vehicle
Vehicle	In a car, taxi, or ride-share
Indoor	Inside a building, venue, or enclosed public space

Day vs. Night Reporting

Time of day fundamentally changes the safety landscape of any location. Juliapp distinguishes between day reports (06:00–20:00) and night reports (20:00–06:00) based on the timestamp of the incident.

Night-time incidents are filtered and weighted separately in the routing algorithm, since safety conditions often differ dramatically after dark.
The filter bar allows users to view reports by time-of-day: morning, afternoon, evening, or night.
Route planning factors in whether the trip is occurring during day or night hours, adjusting danger scores accordingly.

Section 05

Urban Context Categories

Beyond how the reporter was traveling, Juliapp captures where the incident occurred in terms of urban typology. This spatial metadata enables pattern analysis across different types of urban infrastructure.

Context	Description
Street	Public road, sidewalk, or pedestrian path
Parking	Parking lot, underground garage, or parking structure
Store / Mall	Commercial area, shopping center, or retail environment
Metro	Metro station, platform, or underground passage
Bus stop	Bus stop, bus shelter, or transit interchange
Park	Public park, garden, green space, or playground
Restaurant / Bar	Food and drink establishment, terrace, or nightlife venue
Building	Residential building entrance, lobby, hallway, or stairwell
Other	Any context not covered above (user can specify in free text)

Urban context data helps identify spatial safety patterns — for example, whether incidents cluster around transit stations, parking areas, or parks at specific times. This information is used in neighborhood safety scoring and city-level analytics.

Section 06

Community Verification System

Juliapp relies on community participation to validate the accuracy and relevance of safety reports. Multiple mechanisms work together to establish trust.

Confirm & Refute

Any user can confirm or refute an existing report. This crowdsourced validation creates a confirmation score for each pin:

Reports with more confirmations are displayed more prominently and carry greater weight in danger scoring.
Reports with a high refutation ratio may be visually de-emphasized or flagged for review.
Users can only vote once per pin, preventing manipulation.

Identity Verification

Users can optionally verify their identity through a third-party provider (Didit). Verified users receive a visible badge on their profile. Reports from verified users carry additional weight in the confirmation system, as their identity has been independently confirmed.

Impact Score & Milestones

Juliapp tracks each user's positive contribution to community safety through an impact scoring system. Points are earned for:

Creating reports that receive community confirmations
Confirming or refuting other users' reports
Participating in community discussions
Maintaining reporting streaks (consecutive days of activity)
Completing community challenges (weekly community goals)

Milestones are awarded at key thresholds and visible on the user's profile, incentivizing consistent, quality participation.

Section 07

Danger-Aware Routing Algorithm

Juliapp's trip planner does not just find the fastest route — it evaluates each route's safety profile using community-reported data. Here is how the algorithm works.

Route Corridor Analysis

For each candidate route returned by the routing engine, Juliapp samples up to 20 evenly-spaced waypoints along the path. At each waypoint, it checks for unresolved safety reports within a 200-meter radius. This creates a "corridor" of safety awareness around the entire route.

Severity-Weighted Scoring

Not all incidents carry equal weight. The danger score is computed as follows:

Report Type	Weight	Rationale
Standard report	1 point	Base weight for any unresolved incident near the route
Emergency / SOS report	3 points	Emergency reports signal active, high-severity danger

The total danger score is the sum of all weighted incidents detected within the route corridor. Higher scores indicate routes that pass through areas with more reported incidents.

Time Decay & Night Filtering

The algorithm only considers recent, unresolved reports. Resolved incidents are excluded entirely. When traveling at night, the algorithm can optionally filter to only consider night-time reports (incidents that occurred between 22:00 and 06:00), providing time-relevant safety scoring.

Alternative Route Generation

The routing engine (OSRM) generates up to 3 alternative routes for each trip. Each route is independently scored for danger, and users see all options ranked by both travel time and safety. Available transport modes include:

Mode	Routing Engine	Danger Scoring
Walk	OSRM (foot profile)	Full corridor analysis with alternatives
Bike	OSRM (bike profile)	Full corridor analysis with alternatives
Drive	OSRM (driving profile)	Full corridor analysis with alternatives
Transit	IDFM official schedules	No danger rerouting (fixed transit lines)

Why no danger rerouting for transit? Public transit follows fixed routes and schedules. Juliapp cannot suggest alternative metro lines in the same way it suggests alternative walking paths. Instead, transit routes display nearby incidents along the line for awareness.

Section 08

Data Freshness & Lifecycle

Safety data is inherently time-sensitive. A report from 30 minutes ago is far more actionable than one from 3 days ago. Juliapp uses multiple mechanisms to reflect data freshness.

Time-Based Visual Styling

Pins on the map are styled to visually communicate recency. Recent reports appear with stronger visual presence (larger, more opaque, pulsing for emergencies), while older reports are progressively de-emphasized.

Filter Options

Users can filter the map by report age to focus on the most relevant data:

Filter	Shows Reports From	Best For
< 1 hour	Last 60 minutes	Real-time situational awareness
< 6 hours	Last 6 hours	Current-session safety overview
Today	Since midnight	Daily safety picture
Any	All available reports	Pattern analysis and historical context

SOS Alert Expiry

SOS emergency alerts are designed for immediate, time-critical situations. If an SOS is not manually resolved by the user or a responder, it auto-expires after 2 hours. This prevents stale emergency markers from persisting on the map and causing unnecessary alarm.

Historical Data Retention

While the active map focuses on recent data, historical reports are retained in the database for pattern analysis, neighborhood safety scoring, and trend computation. Older reports are anonymized after 12 months in accordance with the privacy policy.

Section 09

Privacy by Design

Juliapp is built with privacy as a foundational principle, not an afterthought. Safety information and personal privacy are not mutually exclusive — here is how they coexist.

Principle	Implementation
No precise home addresses	Juliapp does not store or infer users' home locations. Reports are placed at user-selected map points, not derived from home address data.
Active-only location	GPS location is only accessed when the app is actively in use (map view, trip navigation, SOS). No background tracking occurs outside of active sessions.
Approximate report locations	Reports are placed at the point the user taps on the map, which may be approximate. The system does not snap reports to exact addresses.
No real-name requirement	Users choose a display name freely. Identity verification is entirely optional and does not expose real names to other users.
GDPR-compliant processing	All data processing has a lawful basis (consent, contract, or legitimate interest). Users can access, rectify, export, or delete their data at any time.
EU data residency	All primary data is stored in European Union data centers. Third-party processors comply with GDPR or have adequate safeguards.

For full details on data collection, retention, and your rights, see the Privacy Policy. Privacy.

Section 10

Limitations & Disclaimer

Juliapp is a powerful community tool, but it has important limitations that every user should understand.

Not a replacement for emergency services. Juliapp is a supplementary awareness tool. In case of immediate danger, always call your local emergency number: 112 (EU), 911 (US), 999 (UK), or 15 / 17 / 18 (France).
Community-dependent accuracy. Report accuracy and coverage depend entirely on community participation. Areas with fewer active users will have fewer reports, which does not necessarily mean those areas are safer.
No independent verification. Juliapp does not independently verify individual incident reports. While the community confirmation system helps surface reliable data, individual reports reflect the subjective perception of the reporter.
Indicative safety scores. Danger scores, neighborhood ratings, and route safety assessments are indicative, not absolute guarantees. A low danger score does not guarantee safety, and a high score does not mean danger is certain.
Temporal limitations. Safety conditions change rapidly. A location that was flagged an hour ago may be safe now, and vice versa. Always exercise personal judgment regardless of what the map shows.
Reporting bias. Certain types of incidents may be underreported (e.g., harassment) or overreported (e.g., perceived suspicious activity). Juliapp data reflects community reporting patterns, which may not perfectly mirror actual crime statistics.

Stay safe. Juliapp is designed to augment your situational awareness, not replace it. Trust your instincts, stay alert, and never hesitate to contact emergency services when you feel unsafe.