This project applies Data Analytics and Multi-Attribute Value Theory (MAVT) to optimize car purchasing decisions. The report provides a structured, quantitative approach to evaluating multiple car attributes, ultimately recommending the most suitable vehicle for a family based on comprehensive data analysis.
- Background: A three-member family seeks a reliable, fuel-efficient car for city driving.
- Objective: Utilize MAVT to make an informed car purchase decision by analyzing multiple attributes and constraints.
- Fundamental Objectives: Identified through discussions, the objectives include cost-effectiveness, reliability, performance, space, and environmental impact.
- Attributes: Key attributes analyzed are:
- Price: €16,000 - €14,000
- Fuel Consumption: 10–3 liters/100km
- Repair Costs: €3,000 - €200
- Age: 10–0 years
- Mileage: 120,000 - 20,000 km
- Reliability Ratings: 0–99
- Performance (Horsepower): 90–150 HP
- Acceleration (0–100 km/h): 20–6 seconds
- Braking Distance: 50–10 meters
- Cargo Space: 700–2,000 liters
- Car Type: Binary (Hatchback or Farmer)
- Environmental Impact (CO2 Emissions): 150–95 g/km
- Data Sources: Car listings from Kamux, reliability ratings, repais cost estimators from Caredge.com and J.D. Power.
- Data Analytics:
- Cleaning and Preprocessing: Ensured data accuracy and consistency.
- Normalization: Scaled attributes using linear and non-linear value functions.
- Attribute Scaling: Applied methods like the bisection method for non-linear attributes (e.g., repair costs, mileage).
- SWING Method: Utilized to prioritize attributes by assigning weights based on their impact on the decision-making process.
- Weight Distribution: Emphasized attributes with higher importance to the family’s needs.
- MAV Scores Calculation: Computed for each car alternative.
- Top Recommendations: The analysis revealed “Skoda Octavia – 2017” as the optimal choice with a MAV score of 0.562.
- Ranking: Detailed ranking of all alternatives based on MAV scores for transparency.
- Theoretical Assumptions: Evaluated assumptions like Mutual Preference Independence and Difference Independence.
- Behavioral Biases:
- Generating Objectives and Alternatives: Addressed biases like desirability bias and omission bias.
- Attribute Scaling: Mitigated scaling biases using natural attributes and external data sources.
- Weight Assignment: Used the SWING method to prevent equalizing bias.
- Limitations:
- Data Reliability: Manufacturer’s data vs. real-world conditions.
- Subjectivity in Ratings: Variability and potential biases in reliability ratings.
- Figures and Tables: Detailed illustrations of fundamental objectives, attribute ranges, decision alternatives, and MAV scores.
- Documentation: Comprehensive details on data sources, value functions, and calculation methods.
- Analytical Rigor: Robust data analytics techniques for accurate and reliable decision-making.
- Quantitative Analysis: Clear, numerical evaluation of attributes and alternatives.
- Visualization: Detailed visual representations of data and results.
- Transparency: Extensive documentation of methods and data sources.
- Bias Mitigation: Strategies to identify and reduce biases in the decision-making process.
This project demonstrates the power of integrating data analytics with decision-making theories like MAVT to provide clear, actionable insights in complex scenarios such as car purchasing.
This project was developed as a group assignment for the "Decision Making And Behavior" course at Aalto University. While the design, implementation, and analysis aspects of the project were collaboratively completed by the team, each individual contributed to specific parts of the project.
The work presented here includes contributions from all team members, and the final deliverables reflect a collective effort. When showcasing this project in a personal portfolio, it is important to acknowledge that it was a group effort. The skills demonstrated and the results achieved are representative of the collaborative work undertaken by the team as a whole.