Data Analysis Project - Fuel Price Dynamics (Time Series, SARIMA, Regression)
Budget: $250 – $750 USD
We are seeking a skilled freelance data analyst to undertake a detailed analysis of local fuel prices (gasoline, diesel, and LPG), their relationship with global Brent oil prices, and the impact of local factors including fuel import volumes. The primary objective is to provide actionable insights that will assist a fuel station retailer in strategic decision-making related to pricing strategies, inventory management, and other operational considerations. Cleaned datasets ready for data analysis is available and will be provided by us.
This project is designed to be ongoing on a quarterly basis. The freelancer will receive updated data from us every quarter to refresh the analysis and generate new forecasts for the upcoming quarter, ensuring a continuous partnership and up-to-date insights.
The freelancer is expected to have access to ChatGPT 4 and is encouraged to use the Data Analyst feature at ChatGPT 4 to leverage capabilities for this project. However, it is crucial that any content or insights derived from AI are thoughtfully enhanced and contextualized by human intelligence. Merely submitting unmodified, AI-generated content will not meet our standards and will result in rejection.
Key Deliverables:
1. Time Series Analysis for Local Fuel Prices:
- Identify seasonal patterns to determine the best purchase and sale times.
- Analyze long-term trends and volatility for gasoline, diesel, and LPG prices.
2. Statistical Modeling (SARIMA for Each Fuel Type)
- Split the dataset for each fuel type into training and testing sets.
- Identify optimal SARIMA parameters (p, d, q)×(P, D, Q, s) for each fuel type through iterative testing and evaluation.
- Conduct a grid search for optimal SARIMA parameters and perform model evaluation using metrics like AIC, BIC, and RMSE.
- Use time series cross-validation to assess the performance of the SARIMA models, ensuring robustness and generalizability.
- Evaluate model performance using metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and others relevant to time series forecasting.
- Once the model is trained and evaluated, deploy it to forecast local fuel prices for the upcoming quarter, providing actionable insights for decision-making.
3. Regression Analysis Against Brent Prices (for each fuel type):
- Analyze the time lag between Brent price changes and their impact on local fuel prices, using cross-correlation to identify optimal lag times.
- Develop regression models to quantify the relationship between Brent oil prices and local fuel prices for each fuel type.
- Interpret the coefficients of the regression models to understand the sensitivity of local fuel prices to changes in Brent oil prices.
4. Analysis of Price Changes Against Past Import Data:
- For each fuel type, examine the correlation between the volume of fuel imported and local fuel prices, identifying significant relationships.
- Determine the lag time between changes in import volumes and their effect on local market prices.
5. Interactive Dashboard and Comprehensive Visualization::
Develop comprehensive visual representations (preferred by Power BI), including scatter plots, line graphs, heat maps, and interactive dashboards, to effectively illustrate the various analyses conducted throughout the project. This includes:
- Displaying seasonal patterns, trends, and volatility in local fuel prices (gasoline, diesel, and LPG).
- Visualizing the relationship and lag times between Brent oil price changes and local fuel prices.
- Illustrating the correlation and impact of fuel import volumes on local fuel prices for each fuel type, with emphasis on significant relationships and lag times.
- Showcasing the forecasted prices derived from the SARIMA models, alongside confidence intervals to highlight forecast uncertainty.
- Provide a comprehensive report detailing methodologies, findings, interpretations, and actionable recommendations.
Skills and Tools Required:
Proficiency in data analysis, statistical modeling, and data visualization.
Experience with software tools such as Excel, Power BI, and programming languages like R or Python.
Ability to work with large datasets and perform complex data manipulations.
Project Timeline:
The freelancer is expected to complete the project within 7 days. However, please include your estimated timeline for delivery when applying.
Payment Structure:
Payment for this project is structured around three milestones, corresponding to the completion of the analysis for each fuel type:
First Milestone: Payment upon completion of the Gasoline analysis.
Second Milestone: Payment following the Diesel analysis.
Third Milestone: Final payment after completing the LPG analysis.
IMPORTANT: Placeholder or speculative bids will not be considered. Submit your final bid based on a thorough understanding of the project's requirements.
Dataset samples are attached.
This project is designed to be ongoing on a quarterly basis. The freelancer will receive updated data from us every quarter to refresh the analysis and generate new forecasts for the upcoming quarter, ensuring a continuous partnership and up-to-date insights.
The freelancer is expected to have access to ChatGPT 4 and is encouraged to use the Data Analyst feature at ChatGPT 4 to leverage capabilities for this project. However, it is crucial that any content or insights derived from AI are thoughtfully enhanced and contextualized by human intelligence. Merely submitting unmodified, AI-generated content will not meet our standards and will result in rejection.
Key Deliverables:
1. Time Series Analysis for Local Fuel Prices:
- Identify seasonal patterns to determine the best purchase and sale times.
- Analyze long-term trends and volatility for gasoline, diesel, and LPG prices.
2. Statistical Modeling (SARIMA for Each Fuel Type)
- Split the dataset for each fuel type into training and testing sets.
- Identify optimal SARIMA parameters (p, d, q)×(P, D, Q, s) for each fuel type through iterative testing and evaluation.
- Conduct a grid search for optimal SARIMA parameters and perform model evaluation using metrics like AIC, BIC, and RMSE.
- Use time series cross-validation to assess the performance of the SARIMA models, ensuring robustness and generalizability.
- Evaluate model performance using metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and others relevant to time series forecasting.
- Once the model is trained and evaluated, deploy it to forecast local fuel prices for the upcoming quarter, providing actionable insights for decision-making.
3. Regression Analysis Against Brent Prices (for each fuel type):
- Analyze the time lag between Brent price changes and their impact on local fuel prices, using cross-correlation to identify optimal lag times.
- Develop regression models to quantify the relationship between Brent oil prices and local fuel prices for each fuel type.
- Interpret the coefficients of the regression models to understand the sensitivity of local fuel prices to changes in Brent oil prices.
4. Analysis of Price Changes Against Past Import Data:
- For each fuel type, examine the correlation between the volume of fuel imported and local fuel prices, identifying significant relationships.
- Determine the lag time between changes in import volumes and their effect on local market prices.
5. Interactive Dashboard and Comprehensive Visualization::
Develop comprehensive visual representations (preferred by Power BI), including scatter plots, line graphs, heat maps, and interactive dashboards, to effectively illustrate the various analyses conducted throughout the project. This includes:
- Displaying seasonal patterns, trends, and volatility in local fuel prices (gasoline, diesel, and LPG).
- Visualizing the relationship and lag times between Brent oil price changes and local fuel prices.
- Illustrating the correlation and impact of fuel import volumes on local fuel prices for each fuel type, with emphasis on significant relationships and lag times.
- Showcasing the forecasted prices derived from the SARIMA models, alongside confidence intervals to highlight forecast uncertainty.
- Provide a comprehensive report detailing methodologies, findings, interpretations, and actionable recommendations.
Skills and Tools Required:
Proficiency in data analysis, statistical modeling, and data visualization.
Experience with software tools such as Excel, Power BI, and programming languages like R or Python.
Ability to work with large datasets and perform complex data manipulations.
Project Timeline:
The freelancer is expected to complete the project within 7 days. However, please include your estimated timeline for delivery when applying.
Payment Structure:
Payment for this project is structured around three milestones, corresponding to the completion of the analysis for each fuel type:
First Milestone: Payment upon completion of the Gasoline analysis.
Second Milestone: Payment following the Diesel analysis.
Third Milestone: Final payment after completing the LPG analysis.
IMPORTANT: Placeholder or speculative bids will not be considered. Submit your final bid based on a thorough understanding of the project's requirements.
Dataset samples are attached.