| Step | Tools / Stack | Rationale |
|------|--------------|-----------|
| a. Content Ingestion | • Python scripts (pandas, ffmpeg) to parse PDFs, transcribe audio (Google Cloud Speech‑to‑Text), extract metadata from video (ffprobe).
• Store raw assets in a cloud bucket (e.g., AWS S3 or Google Cloud Storage). | Centralizes the 1.5 GB of raw files, makes them searchable. |
| b. Normalization & Compression | • Text → UTF‑8, stripped of redundant whitespace.
• Audio → AAC 128 kbps, Video → H.264 720p (≈ 1 MB/min).
• Images → WebP (lossless for archival, lossy for UI). | Reduces bandwidth while preserving quality for a smooth front‑end. |
| c. Indexing & Search | • Elasticsearch (or Typesense) with fields: case_id, date, location, keywords, survivor_name (pseudonym), media_urls. | Enables fast filtered queries for the map/timeline. |
| d. Backend API | • Node.js + Express (or FastAPI) exposing REST endpoints:
/cases?year=2022&lat=34.05&lng=-118.25&radius=50
/case/:id
/resources/:location. | Keeps the front‑end lightweight; serves only the needed slices of data. |
| e. Front‑End Framework | • React (Next.js) for server‑side rendering (SEO friendly).
• Mapbox GL JS (or Leaflet with vector tiles) for the map.
• D3.js for the infographic cards. | Provides a responsive, modern UI that works on desktop & mobile. |
| f. Accessibility Layer | • ARIA‑labeled controls, high‑contrast color palette, keyboard‑only navigation, closed‑captioned media. | Ensures compliance with WCAG 2.1 AA – essential for a project about marginalized communities. |
| g. Privacy & Ethics | • All personal identifiers are pseudonymized.
• Consent metadata stored (who agreed to public display, who opted for “private only”).
• Option to “blur” faces/audio for sensitive entries. | Protects survivors while still allowing powerful storytelling. |
During the investigation, prosecutors referenced a digital‑forensic report labeled “1476 MB Work.” This file contained:
The volume of data (1476 MB) was highlighted in court to illustrate the breadth of digital evidence that can be gathered when victims cooperate with law‑enforcement digital‑forensic units. It also underscored how technology can both empower abusers (via surveillance) and aid survivors (via documentation). vannah sterling latina abuse 1476 mb work
| File Type | Approx. Size | Content Overview | |-----------|--------------|------------------| | Text Messages | 200 MB | 3,200+ messages showing threats, manipulation, and attempts to isolate Vannah from family and friends. | | Video Recordings | 800 MB | 45 videos captured on a personal phone and home security cameras, depicting physical altercations and verbal harassment. | | Audio Files | 150 MB | 70 recordings of confrontations, including instances where emergency services were called but not dispatched. | | Official Documents | 226 MB | Police reports, medical records, restraining‑order filings, and court transcripts. | | Support‑Group Testimonies | 100 MB | Anonymous statements from other Latina survivors who experienced similar patterns of abuse. |
How one woman’s courage is turning a massive digital trove into a catalyst for change in the Latino community. | Step | Tools / Stack | Rationale
The server room hummed like a low‑frequency choir, each rack a stanza of whirring fans and blinking LEDs. Vannah pressed her palms against the cold metal of the rack labeled “Legacy Data Lake – Phase III,” feeling the familiar tremor of anticipation that came with every new migration. Somewhere between the smell of ozone and the faint taste of her abuela’s tamarind candy, a warning pinged on her screen: “File 1476 MB – checksum error.” She opened it, and the first line of raw JSON stared back like a secret waiting to be spoken.
| Metric | Tool | Frequency | |--------|------|-----------| | Unique Visitors | Google Analytics 4 (or Plausible for privacy) | Daily | | Average Session Duration | Same | Weekly | | Map Interactions (pins clicked, timeline drags) | Custom event tracking in the front‑end | Real‑time | | Resource Click‑Through (hotline, PDF downloads) | Event tags + conversion funnel | Weekly | | User‑Submitted Tags | CMS admin panel count | Monthly | | Accessibility Audits (color contrast, keyboard navigation) | axe‑core automated testing | Each release | | Survivor Feedback (optional, anonymous survey) | Typeform embed in “Feedback” tab | Quarterly | The volume of data (1476 MB) was highlighted
| Hook | Description | Implementation | |------|-------------|----------------| | “A Day in Vannah’s Life” | A 2‑minute scroll‑triggered video montage (photos, voice‑over, ambient sound) that plays as the user scrolls from the map’s “home” to the timeline. | Use Scrollama to sync video segments with scroll position. | | “Survivor Spotlights” | Rotating carousel of 5‑minute deep‑dive audio interviews (with consent), each paired with a hand‑drawn illustration. | Store audio as streaming MP3; lazy‑load the illustration as the user clicks “Play.” | | “Legal Milestones” | A side panel that appears when the timeline hits years of key legislation (e.g., 1994 Violence Against Women Act). It links to PDF excerpts of the law and to commentary from legal scholars. | Pre‑populate a JSON file with milestone dates; the UI pulls the data when the slider passes the year. | | “Community Response” | A live‑updating feed (Twitter‑style) where NGOs can post updates, upcoming workshops, or crisis‑line hours. | Integrate a headless CMS (Strapi) with a moderated comment endpoint. | | “What If…?” Scenario Builder | A “choose‑your‑own‑path” micro‑simulation where users make decisions (e.g., “Report to police?” “Seek community support?”) and see statistically‑derived outcomes based on the archive data. | Use a simple decision‑tree JSON; outcomes displayed with animated charts. |