An overview of 14 years of buying pattern in Indian automobile industry

Sangamesh K S

March 14, 2018

Today in this article we will pick the dataset from Indian automobile industry and try to look at the overall sales and pattern of Indian automobile industry.

Introduction

The dataset is collected from the government website data.gov.in which is a free dataset provider and if you desire to get the dataset you can visit:https://data.gov.in/catalog/category-wise-automobile-production

With the dataset we will load the dataset, transform, plot and decompose the data. I am not going to cover the predictive modeling, if interested you can go for any regression or time series forecasting for the same.

This article also shows how economic slowdown and recession have impacted the Indian economy and automobile industry. Visa versa how economic growth have contributed to Indian automobile industry

We will load the dataset and look at the head

df<-read.csv("C:/Users/Sangmesh/Downloads/Table-20.6-All_India_SYB2016_1.csv")
head(df)

##                         Category                        Segment X2001.02
## 1       Passenger Vehicles (PVs)                 Passenger Cars   500301
## 2       Passenger Vehicles (PVs)         Multi-Utility Vehicles   169418
## 3       Passenger Vehicles (PVs) Total Passenger Vehicles (PVs)   669719
## 4 Commercial Vehicles - M & HCVs             Passenger Carriers    20283
## 5 Commercial Vehicles - M & HCVs                 Goods Carriers    76469
## 6 Commercial Vehicles - M & HCVs                 Total M & HCVs    96752
##   X2002.03 X2003.04 X2004.05 X2005.06 X2006.07 X2007.08 X2008.09 X2009.10
## 1   557410   782562   960487  1046133  1238032  1426212  1516967  1932620
## 2   165920   206998   249389   263167   307202   351371   321626   424791
## 3   723330   989560  1209876  1309300  1545234  1777583  1838593  2357411
## 4    21156    27628    30419    28982    32828    46542    40995    46026
## 5    99346   138495   184388   190313   261438   248415   151288   204145
## 6   120502   166123   214807   219295   294266   294957   192283   250171
##   X2010.2011 X2011.12 X2012.13 X2013.14 X2014.15
## 1    2453113  2775124  2668633  2519281  2590917
## 2     534183   370945   564928   568692   629255
## 3    2987296  3146069  3233561  3087973  3220172
## 4      54552    54156    50024    41175    49360
## 5     289990   330645   228536   180381   219193
## 6     344542   384801   278560   221556   268553

We will rename the column and remove the row 17 and 19 which are unnecessary

colnames(df)<-c("Category","Segment",2001,2002,2003,2004,2005,2006,2007,2008,2009,2010,2011,2012,2013,2014)
df<-df[-c(17,19),]

Three-wheeler Auto Sales Pattern in India from 2001 to 2014

Now let’s plot the data. We have select the data which is in 17th row and then transpose t() the data, remove the categorical names, convert it to numeric and then the data is good to plot.

auto<-t(df[11,])
auto<-as.numeric(auto[-c(1,2),])
plot(auto,type = "b",xlab = "Year",main = "Yearly sales of Passenger Auto")

We are able to see an upside trend in the auto sales data. Now let’s see is decomposed in time series. For which we will do a small trick we will decompose 2 years into one and plot 2001 to 2014 as 1 to 7 years. As we can’t decompose the dataset for single observation.

library(TTR)
plot(decompose(ts(auto,frequency = 2,start = c(1:14))))

Passanger cab Sales Pattern in India from 2001 to 2014

Now let’s plot the data. We have select the data which is in 7th row and then transpose t() the data, remove the categorical names, convert it to numeric and then the data is good to plot.

cab<-t(df[7,])
cab<-as.numeric(cab[-c(1,2),])
plot(cab,type = "b",xlab = "Year",main = "Yearly sales of Passenger cab")

Now let’s decompose the data as the previous one where we will merge 14 into 7 years

plot(decompose(ts(cab,frequency = 2,start = c(1:14))))

Passanger bike Sales Pattern in India from 2001 to 2014

Now let’s plot the data. We have select the data which is in 14th row and then transpose t() the data, remove the categorical names, convert it to numeric and then the data is good to plot.

bike<-t(df[14,])
bike<-as.numeric(bike[-c(1,2),])
plot(bike,type = "b",xlab = "Year",main = "Yearly sales of Passenger Bike")

Now let`s decompose the data as the previous one where we will merge 14 into 7 years

plot(decompose(ts(bike,frequency = 2,start = c(1:14))))

Passanger bike Sales Pattern in India from 2001 to 2014

Now let’s plot the data. We have select the data which is in 1st row and then transpose t() the data, remove the categorical names, convert it to numeric and then the data is good to plot.

car<-t(df[1,])
car<-as.numeric(car[-c(1,2),])
plot(car,type = "b",xlab = "Year",main = "Yearly sales of Passenger Car")

Now lets decompose the data as the previous one where we will merge 14 into 7 years

plot(decompose(ts(bike,frequency = 2,start = c(1:14))))

Conclusion

We have looked into only 4 segments i.e. Bike, cabs, Auto and car. If you are interested you can try various other segments and try out other combinations like predictive forecasting.

This article was able to demonstrate how coding and analytics can be used for understanding patterns in the time series.

In article also able to find a unique pattern in the data. Wherein when there was slowdown or recession the markets like auto, cab and car have shown a downtrend but two wheeler markets have shown an exponential growth.