TITLE: DESCRIPTIVE ANALYSIS & VISUALIZATION OF

REAL DATA: A STUDY ON ANALYSIS OF CAR SALES IN NORWAY

July 20, 2019

 Table of Contents

1. Automobile Sector................................................................................................................................... 3
1.1 Introduction to Automobile Sector ............................................................................................. 3
1.2 Automobile Sector in Norway...................................................................................................... 3
2. Description of Data .................................................................................................................................. 4
2.1 Source of the Dataset ......................................................................................................................... 4
2.2 Description of fields of Datasets ........................................................................................................ 4
3. The Analysis .............................................................................................................................................. 6
3.1 Models of Central Tendency .............................................................................................................. 6
3.2 Models of Dispersion ......................................................................................................................... 9
3.3 Models of Shape ............................................................................................................................... 11
3.4 Analysis through Probability ............................................................................................................ 12
3.5 Probability Distribution.................................................................................................................... 13
3.6 More Analysis through Visualizations ............................................................................................. 13
4. CONCLUSION .......................................................................................................................................... 16
1. Automobile Sector

This statistical analysis report is based on the sales of variety of cars in Norway from year 2008-
2017. Hence it is apt if we start with the description of the automobile industry from which data
is taken. Here in this section we start with a brief introduction to automobile industry followed
by the present scenario of automobile sector in Norway.

1.1 Introduction to Automobile Sector

The automobile industry comprises of all those companies involved in the manufacture of motor
vehicles, including components such as engines and bodies, but excluding tyres, batteries and
fuels. However, the industry’s principal products are passenger automobiles and light trucks
including vans and sports utility vehicle. Sales of cars depends on many factors like appearance,
culture of the country, brand allegiance of customers, reputation of the make, etc. But for me the
most important factor still must be cost vs efficiency.

1.2 Automobile Sector in Norway

In 2018 the automotive industry in Norway accounted for nearly USD 6.4 billion in sales, and
this industry that has been a well-kept secret is now beginning to get the attention it deserves on
the international level. The reasons for this positive development are numerous.

Opplysningsrådet for Veitrafikken (OFV), Norwegian road association, held a business summit
for its member organizations, where they presented the annual presentation. Among the
highlights for the year, OFV reported all-time-high sales of electric cars, with fully electric and
plug-in hybrid cars accounting for 48.4% of all new car sales (compare to 12% for Sweden and
24% for Denmark). Norway is the leading market for electric vehicles, and it keeps showing the
rest of the world how fast electric vehicle adoption can happen. In 2018, EV sales grew by 40%
in Norway and 1 out of 3 vehicles sold in the market was a zero-emission vehicle. According to
the Norwegian Road Traffic Information (OFV) out of the 147,929 new passenger cars registered
in 2018, 31.2 percent were all-electric vehicles with Nissan's Leaf electric car snapping the title
as the most popular car in Norway. Among 10 most popular cars for the year, OFV reported,
there was only one fossil fuel vehicle.
2. Description of Data

Description of data also known as ‘metadata’ is the most basic and sometimes forgotten aspect of
analysis. Metadata serves variety of purpose, but the most common is it contains information
needed to understand and effectively use the data. In this section we will discuss about the source
of data and descriptions of fields of data so that anyone can understand them.

2.1 Source of the Dataset

The datasets have been exported from a website called Kaggle. Kaggle is an online community
of data scientists and machine learners, owned by Google LLC. Kaggle allows users to find and
publish data sets, explore and build models in a web-based data-science environment, work with
other data scientists and machine learning engineers, and enter competitions to solve data science
challenges.

Kaggle acknowledged that it gathered data from OFV. Opplysningsrådet for Veitrafikken (OFV)
is a politically independent membership organization that works to get politicians and authorities
to build safer and more efficient roads in Norway. The organization has about 60 members,
representing different types of road users. Members are leading players in road safety, car owner
associations, public transportation companies, shippers, car dealers, oil companies, banking,
finance and insurance, road builders and general contractors. Hence, we can say that, for this
statistical analysis study the data used is Secondary Data.

2.2 Description of fields of Datasets

For this statistical analysis, we will analyse from data from two tables. Description of the two
tables and their fields are as follows:

 Monthly summary of top-20 most popular models (by make and model)
- norway_new_car_sales_by_model.csv

Fields Description
Year year of sales
Month month of sales
Make car make (e.g. Volkswagen, Toyota, Tesla)
Model car model (e.g. BMWi3, Tesla S75)
Quantity number of units sold
Pct percent share in monthly total
  Summary stats for car sales in Norway by month - norway_new_car_sales_by_month.csv

Fields Description
Year year of sales
Month month of sales
Quantity total number of units sold
Quantity_YoY change YoY in units
Import total number of units imported (used cars)
Import_YoY change YoY in units
Used total number of units owner changes inside
the country (data available from 2012)
Used_YoY change YoY in units
Avg_CO2 average CO2 emission of all cars sold in a
given month (in g/km)
Bensin_CO2 average CO2 emission of bensin fueled cars
sold in a given month (in g/km)
Diesel_CO2 average CO2 emission of diesel fueled cars
sold in a given month (in g/km)
Quantity_Diesel number of diesel fueled cars sold in the
country in a given month
Diesel_Share share of diesel cars in total sales
(Quantity_Diesel / Quantity)
Diesel_Share_LY share of diesel cars in total sales a year ago
Quantity_Hybrid number of new hybrid cars sold in the country
(both PHEV and BV)
Quantity_Electric number of new electric cars sold in the
country (zero emission vehicles)
Import_Electric number of used electric cars imported to the
country (zero emission vehicles)

Note: The numbers on sales of hybrid and electric cars is unavailable prior to 2011.
3. The Analysis

In statistics, analysis is an approach to analyzing datasets to summarize their main

characteristics, often with visual methods. Analysis and Visualization actually gives insight
about the data beyond modelling and hypothesis testing. In this section we will be using
summarization, graphs and interactive dashboards to present a visual story from this dataset.

3.1 Models of Central Tendency

Beginning our analysis, we can check the total number of cars sold by each manufacturer,
throughout the period of 2007-2016. The results have been presented in the following table:

From the table we can see, that the top 2 car manufacturers are Volkswagen and Toyota, and
they lead the market share by a large margin compared to other.
Of course, one of the factors will be many of the car companies started their operations after
2011 in Norway. But besides that, what makes these companies attract double of the customers
than their competitors.

Let’s look at the following visualization to see which models actually sells the most for Toyota
and Volkswagen.

As we can see, Volkswagen has the best-selling car as Volkswagen Golf, which sold more
models from 2007-2016, than the top two Toyota cars (Yaris and Auris) combined.

Although we can see Volkswagen sold more cars, but the trend in the visualization give a picture
which is something to worry about for Volkswagen. Volkswagen sales are dipping whereas
Toyota sales are growing.

Average: Since few car manufacturers started their business from late 2000’s. It will be unfair to
compare the market share from their total numbers. Let’s look at following visualization of
average car sold, to get a good picture
The average car sales clearly tell Volkswagen sells the most car every year. But Tesla and
Mitsubishi hold the next Ranks and not Toyota, which clearly tells that what we suspected that
years of operations of manufacturers indeed played a role in the numbers

Median: Now let’s look at the below visualization of median car sales, to see that its not late
surge for Tesla and Mitsubishi but their numbers were in fact constant throughout the period.

Median again proves it right what we suspected that Tesla and Mitsubishi were not constant as
they stand below the manufacturers like Toyota and Nissan.
Mode: Lets analyze the fourth central tendency and get the list of the best-selling car from each
make. Following Visualization shows the most frequent sold cars of each make

3.2 Models of Dispersion

In previous section, analysis was done on basis of models of Central tendency and got some
interesting insights. Here in this section we will analyze the data set using Models of Dispersion.

Here, in this section we will calculate the Standard deviation of top 5 Car manufacturers in
Norway and observe the deviation from the mean.

Again, a very disturbing number projected by Volkswagen. The deviation from mean is huge
which is something Volkswagen management should address.
Here lets also observe the variance of those manufacturers using Visualization:

Let’s also visualize the Range and interquartile range of these top 5 Car manufacturers and see
the yearly statistics.
3.3 Models of Shape

Skewness is the measure of extent of symmetry in the data. Kurtosis is the measure of shape of the
distributions tail relative to overall shape. Here in this section we discuss the models of shape –
Skewness and Kurtosis to see if we can get any valid insights.

Here we analyze the skewness and kurtosis of yearly sales of Volvo Model – V40

Following is the data of sales:

Sum of
Row Labels Quantity
2012 450
2013 3833
2014 3247
2015 400
2016 267
Grand Total 8197

Following is the graph of the distribution of sales yearly:

Skewness: 0.446380749
After calculation Skewness turns out to be greater than 0 i.e.: 0.446380749. Hence, we can say
its positively skewed.

Kurtosis: 0.046155831
After calculation Kurtosis turns out to be less than 3. Hence, we can say that it is platykurtic in
nature.

3.4 Analysis through Probability

Inferences can be reached through probability. We can determine the likelihood of car sales, and
outcomes of events. The sales of car event is mutually exclusive (a Toyota car cannot be a Honda car)
and in this section, we use Relative frequency method which uses data gathered from trials of
experiment conducted earlier.

1. Finding probability that cars sold in 2008 is Toyota.

Car sold is Toyota

P (Y=2008 & C= Toyota) = Cars sold in 2008

12503
= 57789 = 0.21

2. Finding Probability that Cars sold monthly, Volkswagen Golf has sold more than 500
units monthly.

Since we have 10 years of data, so 10*12 gives us the sample space

months where volkswagen golf sold > 500

P (Volkswagen Golf monthly sales > 500) =
No. of Time periods

83
= = 0.69
120

3. Using Conditional Probability, Find the cars sold in 2013 is Volvo V40, given the cars
sold is from the make Volvo.

car sold in 2013 is V40

P (car is V40 in 2013| car make is Volvo) =
cars sold of make Volvo

3833
= = 0.37
10290
3.5 Probability Distribution

From year 2007-2016, an average of 5776.3 cars were sold by Ford annually. Modeling this
situation with a Poisson distribution, what is the probability that 3000, 5000, 6000 cars will be
sold annually?

Here λ = 5776.3

5776.3^3000 e^-5776.3
P(X=3000) = = 0.101
3000!

5776.3^4000 e^-5776.3
P(X=4000) = = 0.145
4000!

5776.3^5000 e^-5776.3
P(X=5000) = = 0.1625
5000!

5776.3^6000 e^-5776.3
P(X=6000) = = 0.158
6000!

5776.3^7000 e^-5776.3
P(X=7000) = = 0.128
7000!

Chart Title
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
P(X=3000) P(X=4000) P(X=5000) P(X=6000) P(X=7000)

Poisson distribution follows a normal distribution

3.6 More Analysis through Visualizations

Here, we will investigate the other dataset, and check the buying pattern of customers over the
year and how environment has played an important role in what car customers are buying.
The following Visualization shows the trend sales of Cars and sales of Diesel cars over the year
from 2007-2016

The Visualization clearly shows the lack of interest in buying Diesel cars over the years. And
why is it so? Because the customers are now educated and aware of environmental issues caused
by Diesel cars.

But how true is that Diesel cars are causing pollution. Following Visualization will compare
sales of diesel cars with the average CO2 emission trends.
We can see there is a slight decrease in the CO2 emission as compared to the rapid decrease in
Diesel cars. So, there are more factors like rise in sales of zero CO2 emission cars. Following
visualization will tell us about the Decline of sales of diesel cars and rise in sales of hybrid and
electric cars.

Since we have the data of electric and Hybrid car sales from 2011, But we can see how Hybrid car sales
are growing rapidly and we can see that at the start of 2017, it has overtaken the diesel car in terms of
sale.
4. CONCLUSION

In conclusion, we can clearly infer from our statistical analysis that Customers in Norway are
more eager in buying Hybrid and electric cars than diesel cars. This report clearly indicates what
brands customer are favoring. It also looks at the social aspect and tells us that Diesel cars are
causing harm in environment which has led to the decline in sales over the years.

The most important conclusion to be derived is the report on Volkswagen. This report should act
as a warning bell for Volkswagen group. Their popularity is decreasing year by year. Their sales
are not constant as we can see that the median of sales shows huge variation. The average sales
shows good numbers but only because of one product – Volkswagen Golf. The decreasing sales
is also because of the fact that they don’t have any Electric cars in market.

Lastly, it can be concluded with the fact that if Volkswagen and Toyota don’t act quickly, then
the Norway car market will soon be taken over by BMW and Tesla.