Water Quality Analysis

Isaiah Jenkins

Project Overview

This project involves an Exploratory Data Analysis (EDA) of the California Department of Water Resources water sample dataset, focusing on chemical and physical parameters of water sources across California. The primary goal was to analyze key attributes, such as the 'parameter' (measured constituent) and 'result' (measured value), with a specific focus on pH levels in relation to sample depth. The analysis was conducted using Python, leveraging libraries like Pandas, NumPy, Matplotlib, Seaborn, and SciPy for data processing, statistical analysis, and visualization.

Dataset

The dataset, sourced from the California Department of Water Resources, contains current and historical data on water quality parameters across various counties. This project analyzed a subset of 500,000 records from the CAWaterQualityLabResults.csv file, focusing on:

• Key Attributes:
  • parameter: The constituent measured (e.g., pH, Conductance, Dissolved Boron).
  • result: The measured value of the constituent.
  • sample_depth: Depth at which the sample was collected.
  • Other attributes include station_id, station_type, latitude, longitude, and county_name.
• Scope: The analysis focused on pH levels in Alameda County groundwater samples, comparing pH distributions at maximum and minimum sample depths.

Preprocessing

• Loaded a subset of 500,000 rows to manage computational resources.
• Removed rows with missing values in sample_depth and result to ensure data integrity.
• Converted result to numeric format, handling non-numeric values by coercion to NaN, and dropped remaining NaN values.
• Final dataset size: 187,326 records after cleaning.
• Analyzed pH data by splitting into maximum and minimum sample depth groups for comparison.

Analysis

The EDA included the following steps:

1. Data Exploration:
   • Examined dataset structure, identifying 18 columns with a mix of numeric (e.g., result, sample_depth) and categorical (e.g., parameter, county_name) data types.
   • Identified missing values in sample_depth (29%), result (1.1%), and other columns, addressed through targeted cleaning.
2. Statistical Analysis:
   • Calculated descriptive statistics (e.g., mean pH for maximum depth: 7.896, minimum depth: 8.03).
   • Performed a two-sample t-test to compare pH levels at maximum and minimum depths, yielding a p-value of 0.0483 (below alpha = 0.05), suggesting no significant difference in pH due to sample depth.
3. Visualizations:
   • Generated kernel density estimation (KDE) plots using Seaborn to visualize pH distributions for maximum and minimum sample depths.
   • Used Matplotlib for plotting to inspect data distributions and relationships.

Key Findings

• pH Analysis: The t-test indicated no significant difference in pH levels between maximum and minimum sample depths in Alameda County groundwater (p-value = 0.0483), suggesting sample depth does not strongly influence pH acidity.
• Data Quality: Missing values in sample_depth and result required careful cleaning to ensure reliable analysis.
• Dataset Insight: The dataset's extensive variables provide a robust foundation for water quality analysis, with potential for broader geographic and parameter-based studies.

Installation

To run this project, ensure you have the following dependencies installed:

pip install pandas numpy matplotlib seaborn scipy scikit-learn

Download the CAWaterQualityLabResults.csv dataset from the California Department of Water Resources and place it in the data/ directory.

Usage

1. Clone the repository:

   git clone https://github.com/your-username/water-quality-analysis.git
   cd water-quality-analysis

2. Set up the dataset:

   • Place CAWaterQualityLabResults.csv in the data/ directory.

3. Run the Jupyter Notebook:

   jupyter notebook Water_Quality_Analysis.ipynb

4. Follow the notebook to explore the data, perform statistical analysis, and generate visualizations.

Next Steps

• Expand Geographic Scope: Analyze pH levels across other California counties to identify regional variations.
• Filter by Water Type: Differentiate between groundwater and surface water to assess pH variations by water source.
• Advanced Analysis: Incorporate additional parameters (e.g., Conductance, Dissolved Chloride) and explore correlations using multivariate techniques or machine learning models.