Customer Behavior Analysis for Electric Vehicle Adaption
Introduction:
- As the world aims for a more sustainable future, electric vehicles (EVs) have emerged as an important tool in the fight against climate change. This extensive investigation uses digital twins – virtual duplicates that imitate real-world events to investigate the factors driving electric car adoption. By employing data-driven insights, we reveal the hidden dynamics that impact consumer behavior and direct the industry toward a more educated, strategic approach to EV deployment.
Steps taken:
- Project Planning
- Literature Review
- Data Collection
- Data Preprocessing
- EDA(Exploratory Data Analysis)
- Feature Engineering
- Modelling
- Final Evaluation
Project Planning
Problem Statement:
- As the world progresses toward more sustainable modes of transportation, electric vehicles (EVs) have emerged as a viable alternative to traditional gasoline-powered automobiles.
- However, high prices, limited range, and a lack of charging infrastructure keep preventing widespread EV adoption.
- As a result, understanding consumer behavior surrounding EVs is critical for manufacturers, policymakers, and marketers.
- Traditional methods of investigating consumer behavior, including as surveys and focus groups, are limited in their ability to capture real-time data and dynamic changes in consumer preferences.
Business Needs:
- The goal of this research is to use digital twins to perform a detailed investigation of customer behavior in terms of electric vehicle adoption. The study aims to acquire important insights into the current situation of the electric car market and find potential areas for improvement by analyzing a complete dataset that includes essential parameters such as ‘Postal Code’, ‘Model Year’, ‘Make’, ‘Model’, and ‘Electric Range’.
Data Collection
The dataset encompasses a wide range of features, including ‘Postal Code’, ‘Model Year’, ‘Make’, ‘Model’, ‘Electric Range’, and ‘Base MSRP’. Using these features, the primary objective is to develop predictive models capable of accurately forecasting the ‘Electric Vehicle Type’ that consumers are most likely to choose when making a purchase decision.
Data Preprocessing
EDA(Exploratory Data Analysis):
Exploratory Data Analysis(EDA):
Top Vehicle makes:
- Tesla dominates the market with a substantially higher share than other manufacturers. The data includes popular electric vehicle brands such as Nissan, Chevrolet, Ford, and BMW. This data can help guide targeted marketing techniques and infrastructure development activities based on consumer preferences.
- Newer model years tend to have longer electric ranges, with some exceptions. Over time, electric car range has improved, removing a major barrier to adoption. This trend emphasizes the rapid technological improvements in the electric car industry, which might influence consumer decision-making.
- Tesla’s dominant market share is reflected in the popularity of its Model 3 and Model Y variants. Other popular models are the Nissan LEAF, Chevrolet Bolt EV, Ford Mustang Mach-E, and Volkswagen ID.4.
- This data can help advocacy groups customize their marketing efforts and infrastructure development to the region’s most popular models.
- The ‘Postal Code’ characteristic has no link with other numerical features, indicating that the geographic distribution of electric vehicles is influenced by factors not reflected in the dataset.
- ‘Model Year’ has a moderately negative correlation (-0.48) with ‘Electric Range’, implying that earlier model years have shorter electric ranges.
- ‘Base MSRP’ has a weak positive association (0.11) with ‘Electric Range’, indicating that higher-priced electric vehicles typically have longer ranges.
Modelling:
- Building on the EDA findings, we applied a range of machine learning models to predict the ‘Electric Vehicle Type’ that consumers are most likely to choose based on the available features. The following models were investigated, and their scores are displayed below:
Naïve Bayes:
Accuracy: 0.8963998575685451
ROC AUC Score: 0.819275183987384
Precision: 0.92
Recall: 0.96
F1-Score: 0.94
Logistical Regression:
Accuracy: 0.8342735058752975
ROC AUC Score: 0.6880596785827632
Precision: 0.86
Recall: 0.95
F1-Score: 0.90
Random forest classifier:
Accuracy: 0.9999062950954853
ROC AUC Score: 0.9998155689989119
Precision: 1.00
Recall: 1.00
F1-Score: 1.00
Gradient Boosting Classifier:
Accuracy: 0.9999625180381941
ROC AUC Score: 0.9999449127947211
Precision: 1.00
Recall: 1.00
F1-Score: 1.00
- The Random Forest Classifier and Gradient boosting models scored the best, with R-squared scores better than 0.90 indicating good predictive ability. These models can be used to guide future EV development and infrastructure design, resulting in a more informed and data-driven approach to helping the state transition to sustainable transportation.
Recommendations:
Based on our analysis and conclusions, we propose the following measures for electric car advocacy groups to promote widespread use of electric vehicles:
- Urban Focus: Concentrate efforts on promoting the use of electric vehicles in urban areas, where there are more EVs and the impact on lowering emissions and congestion can be greater.
- Collaboration with Electric Utility Companies: Work with electric utility companies to guarantee that charging infrastructure is in place and to promote the use of renewable energy sources to power electric vehicles, which aligns with sustainability objectives.
- Public Charging Station Advocacy: Promote the establishment of more public charging stations to alleviate range anxiety and improve the convenience of owning an electric vehicle.
- Focused Marketing and Incentive Promotion: Use the data gained from the make, model, and pricing study to create focused marketing campaigns and raise awareness of clean alternative fuel vehicle incentives in locations with low adoption rates.
Conclusion:
- By utilizing the prediction models created in this study and executing the proposed techniques, electric car advocacy groups can effectively aid in the transition to a more sustainable transportation system. The broad use of electric vehicles will not only reduce carbon emissions, but will also help to create a cleaner, more environmentally friendly future.