SenseColombo — Turning sensory chaos into predictable information

Overview

For most visitors, a large shopping mall is just a busy place. For an autistic woman, the uncertainty about stimulus intensity — crowds, noise, harsh lighting — is a barrier that generates anticipatory anxiety and often leads to avoiding the space altogether.

SenseColombo is a concept for a mobile sensory-navigation app designed for Lisbon’s Centro Colombo mall: it shows real-time data about noise levels, light intensity, crowd density and available sensory regulation rooms, and plans routes that respect the user’s sensory profile. The goal is not to manage the user — it’s to give her back control.

This was a research-through-design project: clinical literature on autism in women → analysis of six existing assistive products → traceable design guidelines → a fully specified app concept with information architecture, design system and usage scenario.

Honest scope note: this is a theoretically grounded concept, not an implemented product. It assumes a sensor-equipped mall (IoT / beacon network) and was not yet tested with autistic users — that’s the explicit next step.

Understanding the user

We focused on women with Level 1 ASD (“high-functioning” autism) — a deliberately specific choice, because this group is systematically underserved by both research and products.

Why women specifically? The core clinical picture of autism is the same across genders, but many autistic women present a distinct behavioural profile — the female autism phenotype — built on masking: consciously imitating expressions and gestures, scripting conversations in advance, and constantly monitoring their own behaviour to “pass” as neurotypical. The literature we reviewed documents the cost:

• Diagnosis arrives ~10 years late on average after first contact with mental health services, often preceded by misdiagnoses (borderline personality disorder, anxiety, depression).
• Masking is exhausting: participants in the studies we reviewed described needing entire days of isolated recovery after long social interactions, with elevated risk of autistic burnout.
• Sensory sensitivity compounds mental health risk. Bitsika et al. (2021) found that autistic girls (6–17) with hypersensitivity to sounds, lights and textures showed higher depressive symptoms. Grant et al. (2022), with 973 autistic adults (563 women), found women significantly more likely to develop central sensitivity syndromes — with sensory sensitivity and anxiety mediating the path from autistic traits to physical symptoms.

The design-relevant translation: this user doesn’t need another tool telling her what to do. She needs predictability (to defuse anticipatory anxiety), sensory information (to choose her own path), and an exit (regulation spaces when overload hits anyway).

What already exists

We analysed six assistive products and research prototypes, looking at methodology, results and — most importantly — limitations:

The gap: every product addresses social skills, emotional self-regulation or routine management. None integrates real-time sensory management in public spaces. That’s the space SenseColombo occupies.

From evidence to guidelines

Rather than jumping to screens, we converted the research into a traceable guideline system — every design decision downstream links back to a documented need. Grounded in ISO 9241-210 (human-centred design) and autism-specific design literature, the guidelines are organised in three layers:

1. General principles — clarity, consistency, autonomy, empathy.
2. Autism-specific principles — predictability, sensory comfort, reduced cognitive load, literal and direct communication, user control.
3. Specific guidelines in three categories: Interface & Interaction (cool muted palette, no sudden pop-ups or automatic sounds, sensory personalisation, reversible actions, always show what comes next), Communication & Content (literal language, no irony or metaphor, descriptive icons with text labels), Functionalities & Experience (anonymous participation, sensory classification of places, alternative low-stimulus routes, clearly signposted regulation spaces).

Why this matters as method: guidelines with traceability turn accessibility from a vague intention into testable requirements. When we later chose haptic feedback over sound, that wasn’t taste — it maps directly to the documented sensitivity to unexpected auditory stimuli.

The concept

SenseColombo transforms the mall’s “chaos” into predictable information, on the assumption of a sensor-equipped space (IoT sensors / beacons continuously measuring light, noise and crowd levels).

The information architecture is deliberately minimal — three top-level areas to keep cognitive load down:

• Map — the core. Two modes: search places (see a location’s live sensory conditions before committing to it) and plan route (build the visit from home, before even leaving — the app generates the least stimulating path and re-adjusts it in real time).
• Profile — sensory preferences (sensitivity to noise, light, crowds), visual adjustments (text size, contrast, brightness, dark mode), history and favourites.
• Settings — notifications (silent / gentle vibration only), language, privacy, help.

Each mapped location exposes five data points chosen from the user research: noise level, light intensity, crowd density, nearby regulation spaces, and zones to avoid (construction, events).

Designing the interface

The design system applies the guidelines literally:

• Colour as information, not decoration. A three-colour code lets the user read the environment at a glance, without text: green = comfortable, low stimulus; light blue = moderate, tolerable with attention; grey = highly stimulating, advised to avoid. The palette itself (cool, desaturated) was chosen to minimise visual stimulation.
• Regulation spaces with binary state. An open padlock = available, closed = occupied. Literal, unambiguous, no interpretation required.
• Haptics instead of sound. All alerts arrive as gentle vibration + discreet visual signals — never automatic audio. The user gets critical information privately, without public exposure or auditory startle.
• Typography and icons follow the literal-communication principle: simple type, high contrast, descriptive pictograms always paired with text.
• Personalisation as autonomy. Dark mode, adjustable text, and sensitivity thresholds that shape which routes the app recommends — the app adapts to her, not the other way around.

Walkthrough: Joana’s journey

To stress-test the concept end-to-end, we wrote a full usage scenario: Joana, a young autistic woman who avoids malls, needs a specific book from FNAC on a Saturday.

The scenario covers the complete arc — plan (checking FNAC’s sensory conditions from her sofa, setting her sensitivity profile), navigate (colour-coded corridors; a haptic alert reroutes her around a grey zone: “Highly stimulating zone ahead. We suggest a more comfortable detour”), recover (near the food court, overload begins anyway; the app detects the intense zone and offers a regulation space — one available, padlock open — where she resets), and complete (route recalculated, book bought, exit through the calmest path).

The deliberate design point in this story: the app doesn’t prevent the overload — real environments are unpredictable. It ensures the overload has an exit. Autonomy isn’t the absence of difficulty; it’s having the tools to handle it.

Learnings & next steps

• Specificity is a feature. Designing for “autistic women, Level 1, in this specific mall” produced sharper decisions than designing for “accessibility” ever would. Constraints from real clinical literature beat generic personas.
• Traceability makes accessibility defensible. Every UI decision maps to a guideline, every guideline to a documented need. That chain is what separates evidence-based design from decoration.
• The gap analysis was the pivot. Six case studies later, the opportunity wasn’t “another self-regulation app” — it was the unclaimed intersection of sensory data + physical public space.
• What’s missing — and I know it: validation with actual autistic users. The concept inherits its assumptions from literature, not from fieldwork with the target population. Testing the colour code, the alert thresholds and the regulation-space flow with autistic women is the necessary next step before any implementation claim.