US-Fires-Trends-and-Analysis

Project Overview

This project aims to analyze trends, causes, and impacts of fires in the United States using a dataset obtained from Government's Data site. The analysis includes data cleaning and visualization in Python, querying and extracting insights using PostGreSQL, and creating an interactive dashboard using Power BI.

---

Dataset

Name: US Fires Dataset
Source: National -US- Fire- Occuring-Points
Format: CSV
Description: The dataset includes detailed information about fire events such as location, size, cause, and managing agencies.

---

Steps in the Project

1. Data Cleaning and Preprocessing (Python)

Tools Used: Python (Pandas, NumPy, Matplotlib, Seaborn)
Steps:

1. Load Data: Imported the dataset using pandas.read_csv().
2. Handling Missing Values:
   • Replaced or imputed missing data in key columns like STATCAUSE and OWNERAGENCY and other colomns.
   • Assigned default values for boolean fields with missing entries.
3. Data Type Conversion:
   • Converted columns such as DISCOVERYDATETIME to datetime format.
   • Ensured numeric fields like TOTALACRES and LATDD83 were properly formatted.
4. Exploratory Data Analysis (EDA):
   • Generated summary statistics.
   • Visualized trends and outliers using Matplotlib and Seaborn (e.g., boxplots, line charts).
   • 
   • 
5. Export Cleaned Data:
   • Saved the cleaned dataset to a new CSV file for SQL and Power BI integration.

2. PostGreSQL for Data Analysis

Database: PostgreSQL
Steps:

1. Table Creation:
   • Created a table named us_fires_data with properly defined schema, including primary keys and appropriate data types.
   • Example: OBJECTID SERIAL PRIMARY KEY, TOTALACRES DOUBLE PRECISION.
2. Data Import:
   • Loaded the cleaned CSV data into the PostgreSQL table using the COPY command.
3. Key Queries:

   • Extracted trends in fire size (TOTALACRES) by year and cause.

   • Analyzed the impact of different agencies (OWNERAGENCY) on fire containment.

   • Identified the most common causes of fires by region and time.

   • Evaluated fire frequency and severity by SIZECLASS.

   • 

   • 

4. Advanced Insights:.
   • Created temporary views for complex aggregations.

3. Power BI Dashboard

Tools Used: Power BI
Steps:

1. Data Connection: Imported cleaned data from the CSV file .
2. Slicers: Added slicers for:
   • STATCAUSE
   • OWNERAGENCY
   • FIREYEAR
3. KPIs:
   • Total fire events (Count of OBJECTID).
   • Total area burned (Sum of TOTALACRES).
   • Maximum area burned in a single event.
4. Charts and Visuals:
   • Line chart: Fire trends over years (Count of OBJECTID vs. FIREYEAR).
   • Stacked bar chart: Fires by cause and agency (Count of OBJECTID vs. STATCAUSE).
   • Map visualization: Fire locations using LATDD83 and LONGDD83.
   • Area chart: Total acres burned over time by cause.
5. Interactive Dashboard:
   • Organized visuals with an intuitive layout.
   • Applied consistent color themes (e.g., blue for severity, green for successful containment).

Insights

1. Top Cause of Fires: Lightning was identified as the leading cause of fires.
2. Year with Most Fires: Fire incidents peaked in the year 2000.
3. Size Class Distribution: Fires of size class "A" were most common.

How to Run the Project

1. Python Setup

1. Install the required libraries:

   pip install pandas numpy matplotlib seaborn psycopg2

2. Run the Jupyter Notebook in the notebooks/ folder to clean and preprocess the dataset.

2. PostgreSQL Setup

1. Create a new database and table using the schema defined in sql_queries.sql.
2. Load the cleaned data into the table using the COPY command.

3. Power BI

1. Open the power_bi.pbix file in Power BI Desktop.
2. Refresh the data to connect with the latest cleaned dataset or SQL database.
3. Interact with the dashboard to explore trends and insights.

---

Key Insights

• Trend Analysis: Fires have increased over the years, with significant spikes in certain decades.
• Cause Analysis: Lightning remains the most common cause, while human-related causes are growing.
• Agency Impact: Certain agencies are more effective in managing fires based on fire size and frequency.

---

Contributing

Feel free to open issues or submit pull requests to enhance the project. Suggestions for additional features or insights are welcome!