Data-Driven Analysis of Global EV Adoption

Electric vehicle (EV) adoption globally has rapidly accelerated over the last few years. In this exploration, I dig into the data behind that growth, across various geographies and timeframes.

The dataset (MIT license) is roughly 9,500 rows with various regions and parameters, from 2010 to 2022 (updated annually). We'll chiefly explore sales growth of electric vehicles only (ignoring hybrids) through 2022 and the growth of charging infrastructure networks.

Sales growth over time

We start by getting the data from our spreadsheet into our Pandas DataFrame. From there, we filter our dataset for the powertrains we care about (fully electric, aka BEV) and directly use our spreadsheet to input the country we want to filter by in Python. The dataset also includes some projections, but in this analysis, we only care about historically recorded values, so we filter out the projections and keep historical entries.

# get data from the spreadsheet into our DataFrame 
df = cells((0,0),(7,9542), 'EV Raw Data', first_row_header=True)

# if region cell is empty don't try to filter
if cell(2,2) != '':
    df = df.loc[df['region'] == cell(2,2)]

# filter by selected parameter 
df = df.loc[df['parameter'] == cell(2,1)]
# make this a historical analysis with none of the projections included 
df = df.loc[df['category'] == 'Historical']

We note massive growth starting in 2020. Note a clear trend, with 230,000 EV sales in 2020 in the US, 470,000 in 2021, and a whopping 800,000 in 2022. This growth from 2021 to 2022 is a 70% increase in YoY sales.

The projections from our dataset assume continued growth, expecting 2025 sales in the US to come in close to 3M EV sales.

Globally, we see a similar trend. 2021 to 2022 growth is a little muted relative to the US on a % growth basis, but still tremendous at 58% growth to go from 7.3M sales up from 4.6M sales the year prior.

If you find these charts’ styling attractive, these are basic Plotly line charts with minor layout changes, as shown below.

# title gets edited based on the value of the spreadsheet input at cell(2,1)
fig = px.line(df, x="year", y="value", title=c(2,1) + title_description)

# make chart prettier by removing background and making the margins less aggressive
fig.update_layout(
    plot_bgcolor="White",
    margin=dict(l=30, r=30, t=50, b=50),
)

# displaying chart to the sheet 
fig.show()

Percentage of vehicle sales to EVs

Before digging into market share, a significant bit of context is that China and the US are the overwhelming leaders in overall vehicle sales. The International Organization of Motor Vehicle Manufacturers cites China as the overall vehicle sales leader at over 25M units per year, the US in second at over 15M, and 3rd place Japan is down at over 4M sales per year.

The EV growth from our prior analysis is impressive, but to understand its scope, we must compare its market share relative to all other vehicle types (gas and hybrid). Fortunately, the dataset includes a market share data point, so we don’t need to perform this calculation manually. We note the values again for the United States and globally (“World” in our dataset).

For the US, we note an 8% EV share. Interestingly, the overall EV share % globally is much higher at 14%.

Naturally, it’s time we expanded beyond the US and learned which countries contribute to this 14% share.

Countries with the fastest adoption

Instead of burying our heads in the sand and only looking at the United States and total global sales, we can also do a simple analysis to see which countries have garnered the highest % of their sales as EVs. As we learned earlier, China and the US dominate global vehicle sales, so if the US is doing 8% of its sales to EVs. Globally, the number is closer to 14%, so it’s fair to guess China’s EV sales as a % of their market share is much higher than the US, keeping this global % much higher as the US plays catch-up.

To do our ranked analysis, we use a simple filter to filter down by sales share; we leave all countries and then apply a max() calculation to get the order of the highest EV share by country.

# get our data from spreadsheet into Python 
df = cells((0,0),(7,9542), 'EV Raw Data', first_row_header=True)

# filter by selected parameter of market share
df = df.loc[df['parameter'] == cell(2,1)]
# make this a historical analysis with none of the projections included 
df = df.loc[df['category'] == 'Historical']
# filter just by the latest year available 
df = df.loc[df['year'] == '2022']
# exclude buses, vans, etc and just include cars 
df = df.loc[df['mode'] == 'Cars']

# sort by highest share of market 
df = df.sort_values(by=['value'], ascending=False)

# return dataframe to sheet 
df

There are lots of fascinating insights from this list.

Note that large vehicle sales countries like the US and Japan are dragging down total EV sales, so China pulls the % up with their #1 vehicle sales by country plus a high % of 29% of overall sales as EVs.

We also note countries in Northern Europe whose sales are trending towards EV-only. Northern Europe has regulations that give VAT exemptions to EV producers, aggressive emissions targets, and other regulations that have resulted in this rapid transition to electric vehicles in Northern Europe.

Charging networks

As EV adoption accelerates it’s worth exploring how charging infrastructure is adapting. For context, there are 3 core types of chargers:

• Level 1: Standard residential 120V AC outlet (what you use to plug in every day at-home items; this charging type is very slow, commonly doing 120V @ 14amps = ~1.7kW before efficiency losses for an 80 kW EV battery. This means a minimum of 47 hours for a full charge (realistically more due to efficiency losses/the actual power delivered to the battery), which is a fine enough solution for users who have less commuting and can use public infrastructure as needed.

  
  

• Level 2: Installable in residential applications (electric home dryers usually use 240V) and the most common public infrastructure charging stations; level 2 at home typically is up to 48A, sometimes lower. 240V @ 48A = 11.5 kW peak, realistically delivering less power. With level 2 infrastructure, an 80 kW battery can be charged in 5–9 hours. This is the best home solution as you can get a full charge easily overnight, and for public infrastructure, it allows topping up or fully charging if using prolonged parking. This is not the solution for trips requiring faster charging between stops.

  
  

• Level 3 (DCFC): DC fast charging is a purely commercial option that can’t be installed at home. These chargers get a vehicle from empty to full in under an hour. This is the charging solution that resembles topping up at the pump. This kind of charging has a correspondingly higher cost for convenience. Level 3 chargers range from 50kW to 350kW; at 350kW, an 80 kW battery could be fully charged in less than 20 minutes.

In our dataset, when the data provider refers to slow charging, they’re referring to publicly available Level 2 infrastructure. When referring to fast charging, they refer to Level 3 infrastructure.

We start by visualizing charger growth in the US and China.

We note that growth in China has an extremely impressive growth curve for fast charging, while the US has seen rapid growth for slow chargers and lagging growth for fast charging.

Fast charger growth in 2022 (China vs US)

China fast chargers installed/EV sale in 2022: 760k/4.4M (.17)

US fast chargers installed/EV sale in 2022: 28k/800k (.035)

Slow charger growth in 2022 (China vs US)

China slow chargers installed/EV sale in 2022: 1M/4.4M (.22)

US slow chargers installed/EV sale in 2022: 100k/800k (.125)

Charging experiences vary. In some scenarios, lines or fully occupied chargers may exist. In others, chargers may see very little use. Factors like proximity to significant parking, pricing, etc, have obvious impacts.

On average, chargers are going unused a high percentage of the time. For charging installation to be worth the capital expenditure, installers need to be able to project a usage rate that puts them in profitability on a reasonable timeline (like any investment).

Summary

EV adoption has accelerated in recent years at a rapid pace. As soon as the numbers for 2023 are released we’ll likely see similar growth to prior years. Go ahead and dig into this dataset and glean your own insights; our analysis is naturally incomplete — there’s lots of fields in the dataset we ignored that would provide all sorts of interesting insights (projections, sales of vans/trucks/other interesting vehicle types, etc).

To access the dataset and analysis directly:

https://app.quadratichq.com/file/4251cd97-e14d-4788-baf2-9d161eb10625

Dataset:

https://www.kaggle.com/datasets/padmapiyush/global-electric-vehicle-dataset-2023

Other sources:

Vehicle sales: https://www.oica.net/category/production-statistics/2022-statistics/
Charging info:https://www.transportation.gov/rural/ev/toolkit/ev-basics/charging-speeds