In order to better understand the data, we provide an analysis of the Electric Vehicles registered in Washington State. 

There are two main types of electric vehicles operating: Plug-in Hybrid and Battery Electric vehicles. The consumer market in Washington state is primarily represented by battery electric vehicles (approximately 78% of all electric cars that are registered in the state).

It can be observed that the majority of the cars are registered in Eastern Washington (Greater Seattle Area). Cities with the most electric vehicles registered are Seattle, Bellevue, Kirkland, and Redmond. An additional hotspot for electric vehicles in Washington is Vancouver, WA (5,310). It can be explained by the fact that the city is a part of the metro area of Portland, Oregon. 

We are also looking at the maximum driving ranges of the electric vehicles registered in Washington state. Battery electric cars have much bigger driving ranges with a maximum of 337 miles per charge for some models of Tesla. It also indicates that these types of cars will require charging less often and can go longer without the need for a charging station.

From this dashboard, it can be observed that Tesla is the leading maker of Battery electric vehicles representing 49% of all electric vehicles registered in the state (among both plug-in hybrid and battery electric vehicles) or 65% of all battery electric vehicles. There are almost 70 thousand Teslas registered in Washington, followed by Nissan (13,497) and Chevrolet (12,026).

In this part, we are looking at the charging stations located in the state of Washington. From this dashboard, it can be observed that most charging stations are located in the Greater Seattle area (Eastern Washington) and the city of Vancouver, WA which correlates to the distribution of electric vehicles in the state. The correlation between charging stations and the number of electric vehicles can be observed with R-squared: 0.746896 (City level of granularity). 

Looking at the scatterplots at the bottom of the dashboard, it can be noticed that there is a moderate correlation between the level of education obtained by the general population residing in the cities and the number of electric cars as well as the number of charging stations (P-value < 0.0001). At the same time, there was no significant correlation found between the median household income and the count of cars or stations.

In the forthcoming analysis, we embark on a predictive journey to forecast the growth of two critical components over the next 5 years: electric vehicles (EVs) count and number offuel stations. Utilizing historical data as our guide, we used time series analysis to discern patterns and trends. Time series analysis, a statistical technique that analyzes time-ordered data points, helped us to construct models for forecasting. Specifically, we used two forecasting models — the ARIMA (AutoRegressive Integrated Moving Average)  model, known for capturing trends and cycles, and the Polynomial Model, which can anticipate growth based on past trajectories. 

The graph shows the growth in the number of electric vehicles (EVs) over time, with historical data represented by the blue line and two forecasts for future growth. The red dashed line is an "Extended Polynomial Fit" indicating a prediction of rapid growth, suggesting an exponential increase in EV adoption based on past trends. The green dashed line is an "ARIMA Forecast," which predicts a more conservative increase, suggesting that while the adoption of EVs will continue to rise, it may do so at a more gradual pace. The discrepancy between the two forecasts highlights the uncertainty in predicting future trends, with the polynomial fit being more optimistic, likely assuming current growth trends will continue, while the ARIMA model incorporates a more cautious approach, potentially factoring in market saturation or other limiting factors. For year 2024, ARIMA model predicted the number of EVs approximately 45,338 which continues to grow to 78,377 in 2028.  Whereas, for polynomial model, it forecast the number to go from 38,527 in 2024 to 71,154 in 2028.  

The chart displays historical data and future predictions for the number of fuel stations opened per year. The blue line traces the actual number of stations opened over the years, with a notable surge around 2020 followed by stabilization. The red dashed line represents a forecast using an Extended Polynomial Fit, suggesting an optimistic future where the number of stations would continue to rise sharply. In contrast, the green dashed line, showing an ARIMA Forecast, predicts that the opening of new stations will level off, maintaining a steady rate without further significant increases. This variance between the models may reflect different assumptions about market saturation, economic factors, and industry trends impacting the growth of fuel station infrastructure. For year 2024, ARIMA model forecasted the number of fuel stations approximately 377 which continues to grow to 384 in 2028.  Whereas, for polynomial fit, it predict the number to go from 502 in 2024 to 831 in 2028.  There is a big difference in the predictions for number of fuel stations because of the notable fluctuations in our data as polynomial model does not provide as much insight into the nature of the data and are more prone to overfitting but ARIMA model are particularly good at capturing the autocorrelation in the data and adjusting the forecast based on recent changes and is less prone to overfitting and hence tend to give accurate results for this type of dataset. 

Link to Time Series Analysis Notebook