⚡ Florida Power & Light (FPL) — Correlation Analysis and Simple Linear Regression Analysis
📘 Overview
This project examines how meteorological variables—primarily daily mean temperature and precipitation—influence electrical demand and net generation within Florida Power & Light’s (FPL) service territory. Using simple linear regression models and Pearson correlation analysis, the study evaluates linear relationships across multiple Florida locations from 2019 to 2024. This project was completed as part of a Broward College partnership with the mathematics department, the Broward College Foundation, and FPL. FPL funded the project with a grant.
The analysis was conducted entirely in Excel using the Data Analysis ToolPak for regression modeling and correlation computation.
🗂️ Project Structure
fpl-correlation-analysis-simple-linear-regression/
│
├── data/Files include preprocessed data, data visualizations, and statistical findings
│ ├── GBonilla_FPL_Project_Data_2019_V2.xlsx
│ ├── GBonilla_FPL_Project_Data_2020.xlsx
│ ├── GBonilla_FPL_Project_Data_2021.xlsx
│ ├── GBonilla_FPL_Project_Data_2022.xlsx
│ ├── GBonilla_FPL_Project_Data_2023.xlsx
│ ├── GBonilla_FPL_Project_Data_2024_V2.xlsx
│ ├── GBonilla_FPL_Project_Data_Extra_2019.xlsx
│ ├── GBonilla_FPL_Project_Data_Extra_2024.xlsx
│ ├── GBonilla_FPL_Project_Data_Dictionary.csv
│
│ Excel files can be previewed directly on GitHub or downloaded to open in Excel.
│
├── documentation/
│ ├── GBonilla_FPL_Project_Methodology_and_Workflow.pdf
│
├── presentation/
│ ├── GBonilla_FPL_Project_Presentation.pdf
│
└── README.md
🌐 Data Sources
Electrical Demand & Net Generation U.S. Energy Information Administration (EIA)
Daily time series including:
- Actual electrical demand (MWh)
- Forecasted demand
- Net generation (MWh)
- Datetime stamps
The electrical demand data used in this study were sourced from the EIA database, with a focus on FPL data.
Meteorological Data Florida State University (FSU) Climate Center
Daily observations, including:
- Mean, max, and min temperature
- Precipitation
- Station‑level geographic coverage
Meteorological observations were obtained from the FSU Climate Center.
🔧 Methodology
The analysis uses simple linear regression, modeling one independent variable at a time.
Dependent Variables (Y)
- Daily Electrical Demand
- Daily Net Generation
Independent Variables (X)
- Daily Mean Temperature
- Daily Precipitation
Tools Used
- Excel Data Analysis ToolPak
- Regression
- Pearson correlation coefficient
- Scatter plots with trendlines
- R² for goodness of fit evaluation
📊 Statistical Approach
Correlation Analysis
Pearson’s correlation coefficient (r) quantifies the strength of a linear relationship.
Interpretation:
- r → 1 → strong positive correlation
- r → -1 → strong negative correlation
- r → 0 → no linear correlation
Simple Linear Regression
For each city and year, the model is:
Y = b_0 + b_1X
Where:
- Y = demand or net generation
- X = mean temperature or precipitation
- b₁ = slope
- b₀ = intercept
Goodness of Fit
- R² measures how well the model explains variation in Y.
- Higher R² → better fit.
🗺️ Geographic & Temporal Coverage
Models were created for 10 Florida locations across 2019–2024, including:
- Miami
- Kissimmee
- Hialeah
- Ft. Lauderdale
- Daytona Beach
- Melbourne
- Vero Beach
- Naples
- Miami Beach
- Ft. Lauderdale Beach
🔍 Key Findings
🌡️ 1. Temperature Strongly Predicts Demand and Net Generation
Across all cities and years:
- Strong positive correlations (r ≈ 0.86–0.90)
- High R² values (≈ 0.75–0.81)
- Higher temperatures → higher electrical demand
- Lower temperatures → lower electrical demand
Examples:
Miami 2024 — Demand vs. Mean Temperature
- r = 0.90
- R² = 0.8094
- Regression: y = 9911.6x – 382269
Kissimmee 2019 — Demand vs. Mean Temperature
- r = 0.87
- R² = 0.7726
- Regression: y = 5265.5x – 40031
These results show a strong positive linear correlation between mean temperature and both electrical demand and net generation.
🌧️ 2. Precipitation Has Weak Predictive Power
Models using precipitation as the independent variable showed:
- Very weak correlations (r ≈ -0.03 to 0.08)
- R² values near zero
Example:
Ft. Lauderdale 2019 — Demand vs. Precipitation
- r = -0.03
- R² = 0.001
🌴 3. Coastal & High‑Population Cities Show Stronger Relationships
Cities with:
- Larger populations
- Higher tourism
- Coastal climates
…tended to exhibit stronger correlations and higher R² values.
🧾 Conclusions
- Mean temperature is a strong predictor of both electrical demand and net generation across FPL’s service territory.
- Precipitation is not a meaningful predictor in simple linear regression models.
- Simple linear regression provides clear, interpretable insights, but more advanced models would likely improve accuracy.
Adding additional variables—such as humidity, wind speed, seasonality, or holidays—would increase the model’s explanatory power and improve the R‑squared coefficient.
🚀 Next Steps
Future work may include:
- Multiple linear regression
- Time‑series forecasting
- Feature engineering (e.g., heat index, cooling degree days)
- Model comparison across years and regions