Where are Electric Vehicle Charging Stations Located, and Where are they Needed?
California’s electric vehicle market is rapidly growing. As of 2015, California was estimated to have approximately 250,000 electric vehicles on the road, and California was thought to have about 50% of the nation’s stock of electric vehicles. These numbers are expected to grow as more and more people adopt electric vehicles as a way to mitigate greenhouse gas emissions.
Along with this increase in electric vehicles comes an increasing need for charging infrastructure. Pacific Gas & Electric, California’s largest public utility, proposed installing 7,600 new chargers in its territory, more than doubling the current number of charging stations. California’s other public utilities, San Diego Gas & Electric and Southern California Edison, proposed their own plans as well, with SDG&E planning to build 3,500 new stations and SCE seeking to develop the infrastructure necessary to build 1,500 new stations in their respective territories. 
Installing new charging stations requires an understanding of where charging stations are in demand and where new demand might occur. City centers are obviously attractive places for charging infrastructure, as many electric vehicle owners tend to live in urban and suburban environments such as the Bay Area, the Los Angeles Metro Area, or San Diego. But this can lead to certain areas being ignored, and thus making consumers less likely to consider purchasing an electric vehicle.
In order to see where charging stations are needed, it’s first important to see where charging stations are. To do that, I first created a map (click the link or see below) of counties colored by the total number of charging stations that are located within its borders
california_counties.*, count(california_counties.cartodb_id) as stations
california_ev_stations as stations
california_counties.cartodb_id, california_counties.the_geom, countyfp, geoid, name, namelsad
As is probably expected, we see most of the charging stations concentrated in counties that contain, or are nearby, major metropolitan areas. Los Angeles, Orange and San Diego County have some of the highest numbers of charging stations, while Santa Clara and Alameda have similar numbers in the Bay Area.
This is unsurprising, as most electric vehicles are adopted in and around cities, due to factors such as the short driving range of vehicles, which makes them more useful in compact areas, such as large cities.
However, there are issues here as well. Los Angeles county has far and away the largest number of charging stations, but it is also the most populous county in the entire state, with almost double the population of the next largest county: San Diego. Clearly, in order to see where charging stations are and where charging stations should be, we must correct for this somehow.
namelsad, name, population_estimate_2016, stations, per1000, stations/per1000 as stations_per_1000
pop.*, count(pop.cartodb_id) as stations, population_estimate_2016/1000 as per1000
california_ev_stations as ev
california_counties_w_population as pop
pop.cartodb_id, pop.the_geom_webmercator, population_estimate_2016, namelsad, name
In this map (click the link or look above) I have normalized the charging stations by population, by dividing the number of charging stations by every 1,000 people in each county. The result is an approximation of the number of charging stations available per unit of population in each of the California counties.
Here we can see that larger counties like Los Angeles and San Diego suddenly look much different, hovering at just above the average of 0.17 chargers per 1,000 people statewide. Conversely, many counties that look like they did not have sufficient charging stations in the previous map, such as Mendocino or Humboldt county in the northern part of the state, are seen to actually have relatively high numbers of charging stations per 1,000 people.
Using this map, we can argue that many of the larger counties could actually stand to have a better ratio of chargers per unit of population, such as Los Angeles county. However, we can still do better here.
cartodb_id, the_geom_webmercator, sum, EVs, county, stations, EVs/stations as ratio
rebates.*, count(rebates.cartodb_id) as stations, sum/2500 as EVs
california_ev_stations as stations
rebates_by_county as rebates
rebates.cartodb_id, rebates.the_geom, rebates.sum, rebates.county
In this map (click link or look above) I have estimated the number of electric vehicles in each county, by using data on the amount (in dollars) of rebates claimed by residents in each county. The maximum allowable rebate is $2,500 per purchase of an electric vehicle, so the number of electric vehicles was estimated by summing the total rebates by county and then dividing that sum by $2,500.
Once the number of electric vehicles in each county was estimated, I then divided that number by the number of charging stations in each county, to obtain the ratio of estimated number of EVs to charging stations, to determine which counties had poor ratios of EVs to charging stations.
Seeing this, we get a different story than the previous two maps. Unsurprisingly, counties in the Bay Area, and in the Los Angeles and San Diego area have poor ratios of EVs to charging stations, this is compounded by their average number of charging stations per capita, and in spite of their large numbers of total charging stations.
One interesting thing to note is that Fresno has a particularly bad ratio of EVs to charging stations. This is in addition to its extremely low number of charging stations per capita, and fairly average total number of charging stations as well. This may indicate that Fresno may actually be a prime candidate for more charging station infrastructure, in addition to the more populous major metropolitan areas, as they have few charging stations per person, and a large number of electric vehicles per charging station.
This analysis comes with a few caveats. Mainly that the estimated number of electric vehicles per county is just that, an estimate. Rebates on electric vehicles may not be a perfect indicator of the number of electric vehicles for many reasons.
Lack of knowledge can lead people to not claim their rebates, and California has also recently instituted an income cap, such that consumers with individual incomes of greater than $150,000 may not claim a rebate. This could possibly skew the data towards underestimating the number of electric vehicles purchased in counties where the income is above the state average.
Additionally, the amount of rebate claimed can vary based on whether the vehicle is full electric, vs. partially electric. Partially electric vehicles receive less money in rebates than full electric vehicles. For the sake of this analysis, I assumed that all vehicles purchased were fully electric, and eligible for the full subsidy, but in practice this is not the case. This could potentially underestimate the number of vehicles in counties where a greater proportion of electric vehicle sales were partial electric vehicles.
Despite these caveats, it is clear that as long as the majority of electric vehicles continue to be adopted by people in or around major metropolitan areas the majority of charging stations will be built in those areas, however using techniques like these we can examine where there may be future demand for EV charging infrastructure.
The data used in this project was obtained from these sources:
Alternative Fuels Data Center — http://www.afdc.energy.gov/
American Community Survey — https://www.census.gov/programs-surveys/acs/
Center for Sustainable Energy — https://cleanvehiclerebate.org/eng/rebate-statistics