# RSS/API Data Collection & Daily Reports

Project type: Personal project — Windows workflow automation

Role: AI-assisted project development and workflow iteration

Suggested skills: Scripting & Automation, Data Extraction, API Integration, CSV, PowerShell. Use matching tags offered by the platform.

## Short description

A Windows automation project that organizes public RSS/API records into reviewable daily reports. It includes source adapters, keyword scoring, history deduplication and CSV/Markdown output. An offline check of the original core functions turned three synthetic RSS records into two unique report rows. Live data sources, scheduling and notifications were not retested for this portfolio.

## Client problem

Checking multiple public information feeds and copying relevant records into a spreadsheet is repetitive. A useful handoff needs original links, consistent columns and a readable summary.

## Implementation

The PowerShell source contains RSS and several public job-feed adapters. Records are normalized, scored with configurable keywords, flagged by review terms, merged into a history and exported as CSV and Markdown. JSON stores the history. This is deterministic automation; the demonstrated scoring does not call a language model.

## Evidence

- Original source functions were executed in an isolated offline test scope.
- Three synthetic RSS records included one repeated URL; merging produced two records.
- Repeating the merge kept the history at two records.
- CSV and Markdown outputs were generated by the original report function.
- Ten focused checks passed. These are fixture results, not a live-service reliability benchmark.

See `../samples/report-input.xml`, `../samples/report-output.csv`, `../samples/report-output.md` and `../samples/report-history.json`. All sample records are fictional.

## Appropriate small engagement

Adapt one agreed RSS/API source to a specified output schema, add a clearly defined filter, or improve a reporting script. Start with an agreed sample and expected output before expanding to more sources.

## Scope and limitations

The original project uses Chinese report labels. An English client-facing report is a separate adaptation. This portfolio does not claim production deployment, measured time savings, continuous uptime, live source availability, an automated Upwork scraper, or working client notification delivery. Notifications and startup tasks were not executed during this review.

Development used Codex assistance. This is a personal project, not a paid-client case study.