Longread translated automatically
big

Concept

big
Original size 1440x810

Data Sources for Media

Primary sources: interviews with sorting center employees, garbage truck drivers, and recycling plant engineers.

Official Documentation: Regional waste management operator reports (solid municipal waste), recycling GOST standards.

Open Databases: Housing and Communal Services GIS, recyclables drop-off point registries, Recyclemap.

Science and popular publications, and industry reports: articles on new recycling technologies, and press releases about brands' environmental initiatives.

Human and Machine Processing

What belongs to machine processing?

Data Collection: Parsing news feeds, Telegram channels of regional operators, and eco-project websites using keywords such as «recycling, ” „disposal, ” and „secondary raw materials.“

Array Processing: Leveraging LLMs (neural networks) to extract key figures and facts from bulky, multi-page departmental reports.

Analytics: Collecting audience engagement metrics via TGStat and VK parsers.

What will the human do?

Storytelling and Editing: Transforming dry facts and «machine-generated» summaries into vibrant text that resonates with the target audience (city residents and curious minds).

Fact-Checking: Verifying data regarding «greenwashing» (when companies lie about their environmental practices) and cross-referencing automated summaries with reality.

Fieldwork: Organizing shoots, creating on-site reports, and conducting interviews.

Guidelines for Automated Services

For news parsers (RSS/Scrapers):

Keywords for filtering: gather news only that specifically mentions processes (e.g., «plastic shredder,» «sorting line,» «textile recycling»). Ignore general press releases about «nature protection» without substance.

For text-based neural networks (LLMs — summarization):

Prompt Guideline: «Analyze the attached waste disposal report. Extract only: 1) precise volume figures, 2) descriptions of new technologies, 3) logistical chains. Ignore fluff and corporate accolades. Output format: bulleted list.»

For visual neural networks (generating covers/diagrams):

Generating a «magical» quality (sprouts in hands, etc.) is prohibited. Images should focus on an industrial aesthetic—containers, cardboard presses, conveyor belts.

КУДА ЭТО ДЕВАЕТСЯ? Схема данных с разделенем на человеческую и машинную обработку
Project created at 08.11.2026
designschool.am
designschool.am — цифровая экосистема, объединяющая профессиональное портфолио для дизайнеров — студентов, выпускников и сотрудников программы «Коммуникационный дизайн» (совместный проект Российско-Армянского Университета и Школы дизайна НИУ ВШЭ), образовательную онлайн-платформу, площадку, агрегирующую международные онлайн-конкурсы в креативных индустриях и другие профессиональные инструменты.... Show more
We use cookies to improve the operation of the website and to enhance its usability. More detailed information on the use of cookies can be fo...
Show more