srajankulal27-aiml/fake-social-media-account-detection

How srajankulal27-aiml/fake-social-media-account-detection works

This is a FastAPI-backed fake social-account risk classifier. Users submit one profile, paste profile text, upload a screenshot, or upload a CSV; the service parses or validates inputs, runs feature-engineered model inference, and returns labels, probabilities, risk tiers, and (for predictions) feature contributions. The repository also includes offline data preparation, model training and evaluation, dashboards, and SQLite prediction history. Some dashboard and batch wiring is documented but its implementation was not sampled, so those links rely on README statements.

Main entry points: main.py API and pages · predict.py Inference and attribution · config.py Runtime configuration · train.py Model training

Components 20

Web application

API and pages app_api/main.py
Input schemas app_api/schemas.py
Quick-fill parsing app_api/quickfill.py
Screenshot extraction app_api/screenshot.py
Browser interactions app_api/static/js/app.js
Server-rendered views app_api/templates

ML pipeline

Raw data loading src/data_loader.py
Data cleaning src/preprocessing.py
Feature engineering src/features.py
Model training src/train.py
Model evaluation src/evaluate.py
Inference and attribution src/predict.py
Runtime configuration src/config.py
Trained model models/best_model.joblib

Audit and results

Performance dashboard app_api/templates/performance.html
EDA dashboard app_api/templates/eda.html
Prediction history src/database.py
Risk results app_api/templates/predict.html
Batch results app_api/templates/batch.html

Other components

Analyst

Connections 25

Analyst → Browser interactions: enters profile README.md
Browser interactions → API and pages: submits requests README.md
API and pages → Input schemas: validates input app_api/main.py
API and pages → Quick-fill parsing: uses parser app_api/main.py
API and pages → Screenshot extraction: uses extraction app_api/main.py
API and pages → Inference and attribution: runs inference app_api/main.py
Inference and attribution → Feature engineering: derives features src/predict.py
Inference and attribution → Trained model: loads pipeline src/predict.py
Inference and attribution → Risk results: returns prediction README.md
API and pages → Batch results: serves batch flow README.md
API and pages → Server-rendered views: renders pages README.md
Browser interactions → API and pages: requests extraction README.md
API and pages → Prediction history: reads and clears app_api/main.py
API and pages → Prediction history: logs predictions README.md
Raw data loading → Data cleaning: prepares records
Data cleaning → Model training: cleans training data
Model training → Feature engineering: engineers features src/train.py
Model training → Trained model: saves selected model src/train.py
Model evaluation → Performance dashboard: supplies metrics README.md
API and pages → Performance dashboard: serves dashboard README.md
API and pages → EDA dashboard: serves dashboard README.md
Runtime configuration → Raw data loading: sets data paths src/data_loader.py
Runtime configuration → Data cleaning: sets split settings src/preprocessing.py
Runtime configuration → Model training: sets model settings src/train.py
Runtime configuration → Inference and attribution: sets model path src/predict.py

Drawn by GitDiagram on November 14, 2026 from the repository's file tree, README and source files. Read as Markdown ·