🕷️ SpiderNet – Detect. Connect. Predict.

Project Overview

SpiderNet is an intelligent crime analysis and prediction system inspired by how spiders detect vibrations in their webs and sense disturbances. Instead of reacting after crimes occur, SpiderNet focuses on early detection, pattern recognition, and predictive insights using real-world crime data. The system transforms static crime datasets into a dynamic, interconnected intelligence network, enabling better decision-making and faster response.

Problem Statement

Crime systems are reactive, not predictive → action starts after damage is done Rising urban crime is worsened by delayed detection and manual analysis Emerging hotspots go unnoticed until they escalate No real-time intelligence → slow response, poor resource use Dashboards show data, not connections → patterns stay hidden Missed patterns today become crimes tomorrow

Proposed Solution

Spiders sense danger through vibrations in their web — not by sight, but by signal pulses This allows instant identification of threats and precise response SpiderNet adapts this mechanism to crime data by detecting: Crime spikes Pattern clusters Hotspot intensity By analyzing these signals collectively, the system identifies patterns, emerging threats, and hidden connections in real time. Result: Instead of static reports, SpiderNet provides a living, responsive crime ecosystem — just like a spider’s web. A Streamlit-based crime analysis platform for San Francisco PD crime data.
Built for the DataWeb Hackathon, SpiderNet helps law enforcement intelligence teams detect anomalies, hotspots, and patterns in crime data.

---

Features

Tab	Description
📈 Weekly Spikes	Z-score–based anomaly detection on weekly crime volumes
🗺️ District Map	Interactive Folium map with pulse-ring animations and "crime web" spider lines
📊 Breakdown	Category, day-of-week, and hour-level crime breakdowns
🤖 Clusters	KMeans geographical hotspot clustering
🔍 Explainer	Actionable intelligence panel with recommended patrol actions

---

Setup

1. Clone the repo

git clone https://github.com/shagunpathak1508/SpiderNet.git
cd SpiderNet

2. Install dependencies

pip install -r requirements.txt

3. Add the dataset

Download the San Francisco Crime Classification dataset from Kaggle and place train.csv in the data/ folder:

data/
  train.csv

⚠️ The data files are not included in this repo due to their large size (~200 MB).

4. Run the app

streamlit run app.py

---

📁 Project Structure

SpiderNet/
├── app.py                          # Main Streamlit dashboard
├── requirements.txt                # Python dependencies
├── notebooks/
│   ├── SpiderNet_Crime_Analysis.ipynb   # Full EDA notebook
│   ├── ArachneX_analysis.ipynb          # Supplementary analysis
│   ├── cluster_map.png
│   ├── district_ranking.png
│   ├── heatmap_day_hour.png
│   ├── weekly_crime_dashboard.png
│   └── ...
└── data/                           # (not tracked — add your CSVs here)
    └── train.csv

---

Tech Stack

• Streamlit — Dashboard framework
• Pandas / NumPy — Data wrangling
• Matplotlib / Seaborn — Static charting
• Folium + streamlit-folium — Interactive maps
• scikit-learn — KMeans clustering
• SciPy — Z-score statistical anomaly detection

---

Dataset

SF Crime Classification — Kaggle
San Francisco Police Department incident records from 2003–2015.

---

Built with ❤️ for the DataWeb Ideathon