Daily animal control incidents present complex challenges related to public safety, location-based risk, and responsible pet ownership. This project executes a full, end-to-end data analytics pipeline to analyze dog bite incident logs, uncover critical patterns, and deliver insights via dual reporting platforms. It demonstrates expertise in ETL, feature engineering, statistical analysis, interactive web dashboarding (Streamlit), and professional business reporting (Power BI).

Specifically, I drove the entire pipeline:

• Data Transformation (Python/Pandas): Achieved robust data cleaning and imputation using Python with the pandas and numpy libraries, handling mixed data types and standardizing over 76,000 records.
• Feature Engineering: Engineered crucial temporal features (e.g., Time of Day, Day of Week), created demographic bins (e.g., Victim Age Group), and calculated the Report Delay Days metric to assess administrative efficiency.
• Visualization & Analysis: Employed Plotly for creating interactive, web-ready visualizations for trend analysis, and Seaborn for static exploratory charts.
• Dual-Platform Deployment: Presented final outputs as an Interactive Streamlit Application for real-time exploration and a Professional Power BI Report using DAX for sophisticated proportional metrics, optimized for high-quality PDF export.

👉 Click here to explore the deployed dashboard

This view highlights the dynamic, interactive nature of the web application, featuring the collapsed filters and Plotly visuals.

This represents the clean, static, multi-page layout optimized for print, demonstrating DAX measures and sophisticated data visualization.

• 🧹 Robust Data Cleaning (Python/Pandas): Handled mixed data types, imputed missing age and location values, and standardized text categories (e.g., "UNKNOWN").
• ⚙️ Feature Engineering: Created essential columns like Day of Week, Time of Day, Victim Age Group, and Report Delay Days.
• 🗺️ Geographical & Temporal Analysis: Identified peak incident hours (heatmap) and analyzed regional hotspots (city ranking).
• 📊 Dual Visualization Platforms:
  • Interactive Streamlit Dashboard: Multi-page web application featuring Plotly visuals, dynamic collapsible filters, and KPI tracking.
  • Professional Power BI Report: Multi-page report optimized for print (PDF export), utilizing DAX measures and advanced proportional analysis (100% stacked charts).
• 📦 Clean Data Export: Finalized, clean data exported to Excel (.xlsx) with underscore-separated column names for immediate consumption by Power BI.

This project uncovered several actionable findings crucial for public safety resource allocation:

• ⏰ Peak Activity Hours:
  Incidents show a marked spike during late afternoons and evenings (4 PM – 8 PM) and peak significantly on weekends, suggesting higher incidence during leisure time.
• 🧒 Vulnerable Groups:
  The Child (6–12) and Teen (13–17) age groups show a disproportionately high count of incidents relative to other demographics.
• ⚠️ Severity Correlation:
  Bites involving STRANGER relationships show a different proportional severity profile compared to those involving OWNER/VIC, indicating that the context of the incident directly impacts outcome severity.
• 🕒 Reporting Efficiency:
  The Avg Report Delay Days metric highlights administrative lag times, which is crucial for assessing public health response efficiency.

• 📂 Data Load & Clean:
  Raw data (Dog_Bite_Dataset.csv) is loaded.
  Python performs conversion, imputation, and feature engineering (documented in preprocessing.ipynb).
• 📊 Analysis & Visualization:
  Key trends (Time, Age, Circumstance) are discovered and visualized using Plotly in the Streamlit app (documented in eda.ipynb).
• 📤 Data Export:
  The cleaned and enriched dataset is exported to Cleaned_Dog_Bite_Dataset.csv.
• 🖥️ Dashboard Build:
  The Streamlit app (streamlit_app.py) runs the entire cleaning logic once (using @st.cache_data) and presents the interactive, filtered views.
• 📈 Power BI Modeling:
  The CSV file is imported into Power BI for modeling, DAX calculation, and report creation (report saved as powerbi/Report.pdf).

├── 📁 powerbi/
│   ├── Power BI Report.pbix
│   └── Report.pdf
│
├── 📁 imgs/
│   ├── incident_trend_by_year.png
│   ├── incidents_by_day_of_week.png
│   ├── severity_by_relationship_stacked.png
│   ├── top_10_bite_circumstances.png
│   ├── report_page1.jpg
│   ├── report_page2.jpg
│   ├── report_page3.jpg
│   ├── dashboard_page1.png
│   ├── dashboard_page2.png
│   └── dashboard_page3.png
│
├── Dog_Bite_Dataset.csv
├── Preprocessed_Dog_Bite_Dataset.png
├── Cleaned_Dog_Bite_Dataset.png
│
├── preprocessing.ipynb
├── eda.ipynb
├── streamlit_app.py
│
├── requirements.txt
└── README.md

• 🏥 Public Health & Safety:
  Pinpointing high-risk times and areas requiring increased animal control patrols or public education campaigns.
• 🏘️ Community Risk Assessment:
  Allowing neighborhood leaders to identify and mitigate high-frequency risk factors (e.g., proportional risk assessment by victim relationship).
• 💰 Resource Allocation:
  Providing metrics on total treatment cost and report delay for budgetary and administrative planning.
• 💼 Data Science Showcase:
  A powerful demonstration of end-to-end data processing, feature engineering, custom data modeling, and multi-platform dashboarding skills.

1. Clone the repository

   git clone https://github.com/Anushka-Sharma-008/DogBiteAnalysis.git
   cd DogBiteAnalysis

2. Launch Streamlit Dashboard

   streamlit run streamlit_app.py
   

Anushka Sharma
🌐 LinkedIn • 🐱 GitHub 🎓 Learning Data Science, Analytics & Machine Learning

If you found this project helpful or inspiring:

• ⭐ Star this repository
• 🛠️ Fork it to build upon or adapt it for your own use
• 💬 Share feedback or suggestions via Issues/Discussions