Project Title: Statistical Analysis of Literacy Rates in

India Using State-wise Socio-Economic Data

ABSTRACT
This project analyzes how literacy rates in India are
affected by socio-economic factors like sex ratio,
population density, poverty, and unemployment using
state-wise data over several decades. It uses graphs to
show trends and applies a multiple linear regression
model to understand which factors most influence
literacy. The study also reviews government programs
for improving literacy and compares India’s literacy
rates with neighboring countries to find useful
strategies. The goal is to provide insights that can help
improve education and policy planning in India.

INTRODUCTION
“Education is the most powerful weapon with which you
can change the world.”

-Nelson Mandela.

This project emphasizes the critical role of literacy in

India’s social and economic development. It explores
how literacy empowers individuals and contributes to
national growth. Given India's large population and
diversity, improving literacy is essential for achieving
inclusive progress. The study examines how various
factors—like sex ratio, poverty, unemployment, and
urban-rural differences—affect literacy rates over time.
Using statistical tools and visualizations, it aims to
uncover key trends, assess government policies,
compare India’s literacy with neighboring countries,
and provide actionable recommendations to improve
education across the nation.
This project uses data visualization,
statistical modeling, and policy analysis to explore
literacy trends in India. Graphs and charts illustrate
how literacy rates have changed over time and relate
to factors like sex ratio, poverty, and unemployment. A
multiple linear regression model identifies key
influences on literacy. The study also reviews
government initiatives, evaluates their impact, and
compares India’s literacy performance with neighboring
countries to draw useful lessons and recommendations.

METHODOLOGY
Visualization:
 Pie Chart: Shows proportions by dividing a circle
into slices based on the quantity each category
represents.
 Bar Graph: Uses rectangular bars to compare
values across different categories.
 Line Diagram: A graph that uses the point
connected by lines to show trends over time. Ideal
for time series data.
 Multiple Line Diagram: A graph with several lines ,
each representing different dataset , to compare
trends across multiple variable.
 Scatter Plot: A graph of plotted point showing the
relationship between two variables . Each point
representing an observation.
  QQ Plot: A Quantile-Quantile plot compares the
distribution of a dataset with a theoretical
distribution. It helps check normality.
Metrics: Here We use different types of error metrics
(eg. Mae, mse, rmse, mape .
 R² (R-squared): Shows the proportion of variance in
the dependent variable explained by the model.
Ranges from 0 to 1.
 Adjusted R²:Adjusts R² for the number of predictors
to avoid overfitting with too many variables.
 P-value:Indicates the statistical significance of model
coefficients. A lower p-value suggests stronger
evidence against the null hypothesis.
 Correlation:Measures the strength and direction of a
relationship between two variables. Ranges from -1
to +1.
Regression Models:
 Simple Linear Regression:
Models the relationship between a dependent
variable Y and a single independent variable X as:
Y =a+bX
where a is the intercept and b is the slope. Includes
an error term to account for deviations in real data.
 Multiple Linear Regression:
Explores the impact of multiple predictors on a

Y^=b0+b1X1+b2X2+⋯+bpXp
single outcome:

Each coefficient bi represents the effect of an

independent variable Xi on Y, holding all other
variables constant.
EXPLANATORY DATA ANALYSIS

VISUALIZTION
Intially , The study focuses on five socio-economic
indicators: literacy rate, sex ratio, population density,
poverty rate, and unemployment rate. Census-based
data (literacy, sex ratio, population density) are
analyzed by census years, while economic indicators
(poverty, unemployment) are studied by economic
years. Visual tools are used to better understand the
trends.
 Literacy Rate:
Literacy is a vital indicator of socio-economic
progress and human capital. Despite improvements
in India, disparities persist, making it crucial to
understand key factors and assess government
initiatives.
STATE-WISE LITERACY RATE OVER THE
DECADES
1961 1971 1981 1991 2001 2011
80
40
0

Our data spans 50 years, capturing long-term literacy

trends across all Indian states and union territories.A
consistent upward trend reflects improved access to
education and the impact of evolving educational
policies
 Sex Ratio:
Sex ratio, a key demographic indicator, influences
literacy and socio-economic development. A balanced
ratio supports gender equality and education, while a
skewed ratio hinders literacy and growth. According to
2011 census report the sex ratio is given by 940, i.e.
there are 940 females per 1000 males.

1400 STATEWISE SEX RATIO IN INDIA

1200
1000
800
600
400
200
0

1961 1971 1981 1991 2001 2011

Over 50 years of data reveal gradual improvements

in sex ratios across India, with Kerala consistently
leading due to strong social policies. In contrast,
states like Haryana and Punjab lag due to cultural
biases, though awareness and policy efforts have
driven progress nationwide.

 Population Density:
Population density impacts resource distribution and
access to education, influencing literacy rates. High
density may boost educational facilities, while low
density can limit access, affecting socio-economic
development.
 STATE-WISE POPULATION DENSITY OVER THE
DECADES

1961 1971 1981 1991 2001 2011

12000
8000
4000
0

According to the 2011 Census, India’s population

density rose to 382 persons/km² from 324 in 2001,
reflecting rapid urbanization. States like Bihar, West
Bengal, and Delhi show high densities due to fertile
land and urban growth. In contrast, Arunachal Pradesh
and Mizoram have low densities due to difficult terrain
and forest cover. Rajasthan also has a low density,
largely due to its desert landscape. Overall, the data
from 1961 to 2011 highlights significant and varied
population growth across regions.
 Proverty Rate:
The proverty rate is crucial socio-economically as it
directly impacts literacy rates. High proverty restricts
access to education due to financial constrains, leading
to lower literacy levels.
 STATE-WISE PROVERTY RATE OVER THE
DECADES
1993-1994 1999-00 2004-05 2009-10 2011-12
60
30
0
T sh r a ir a sh r l hi ep
a
ha Go an hm tak m ha an du nd ga arh
E /U
d e i y d e ip u ra is t h a a n g el e
T a B ri s a ra n izo d s iln akh Be i D w
A r a a r n a O ja d d
ST al
P H K
Ka ya
p M M
Ra Ta tta
m r t an ha
h &
h es Ch ks
u U W a
ac m ad L
r un am M
A J

States like Bihar, Odisha, and Chhattisgarh have

historically high poverty rates due to low agricultural
productivity and poor infrastructure. In contrast, Kerala
and Punjab show lower poverty levels thanks to better
social indicators and development. Poverty data
reveals significant regional disparities but also positive
trends in reduction. Focused policies are needed for
high-poverty areas. Overall, national poverty decline
reflects economic growth and government welfare
efforts.
 Unemployment Rate:
The unemployment rate is critical in socio-economic
contexts as it affects economic stability and growth.
High employment can limit access to education due to
financial hardships, lowering literacy growth.
 STATE-WISE UNEMPLOYMENT RATE IN
INDIA
UNEMPLOYMENT RATE PER

300
200
100
0
r r i
sh am arh oa ana mi aka esh pu am sha an du ura esh nds vel akh rry
1000

e G i r h a
s
ad As tisg y h t d n zo d i t N rip ad sla Ha ad che
Pr t ar Kas rna Pra Ma i O jas il T Pr r I r L u
H M a
r a h a & K a a
y R
T a m
ar
b a
aga ud
h t P
dh C u h t co N
An
m ad U i
N a&
m M d
Ja r
an ad
STATES/UT n D
a
am
d
An
1993-94 1999-00 2004-05 2009-10 2011-12

As of the above data shows Unemployment rate in

India over the states from year 1993-94 to 2011-12.
Here each bar shows number of unemployed per 1000
person. We can see that Kerala had one of the highest
unemployment rates, particularly in urbun areas and
among educated individuals. Gujarat reported one of
the lowest unemployment rates , with rural and urbun
areas showing minimum unemployment. Tripura also
exhibited high unemployment rates, notably higher in
rural areas compared to the national average. Bihar
and Uttar Pradesh had moderate unemployment rates ,
with notable differences between rural and urban
regions.

MODEL MAKING:
We want to understand the relationship between
literacy rates and sex ratio, population density,
proverty rate and unemployment rate. A multiple linear
regression model is fitted on the data. To do so we use
the State-wise literacy rate, sex ratio, Population
Density, Proverty Rate and Unemployment Rate for
2011 . As the columns , since census of 2021 got
postponded due to Covid lock down and there was
presence of missing data for other years. Each row
represent a different state of union territory . There are
total 35 rows and 5 columns. There is no missing data
for the year 2011.
Our variables are:

Y= Literacy Rates

X1= Sex Ratio

X2= Population Density

X3= Proverty Rate

X4= Unemployment Rate

Here, Y is dependent variable and X1,X2,X3,X4 are

independent variables . Each values of Y, X 1, X2 ,X3 , X4
represents values of the corresponding variables for an
individual State.
First we obtain the scatter plot matrix which is
as follows:
Model 1:
The multiple Regression model is:
Y= 64.5218970+ 0.0171450X1 + 0.0010833X2 - 0.1631505X3 -
0.0203367X4

The efficient of all the variable are not significant to the

model and adjusted R2= 0.4154 , which means 41.54 %
variation is explained by the model. Hence we try to
check multicollinearity and homoscedasticity in the
Variable.
The Correlation Matrix is given by:
Literacy.Rate Sex.Ratio Population.Density Proverty.Rate
Unemployment.Rate

Literacy.Rate 1.00000000 -0.01931145 0.3064560 -

0.2385548 -0.1371591

Sex.Ratio -0.01931145 1.00000000 -0.3143315

0.3323161 0.1189291

Population.Density 0.30645604 -0.31433146 1.0000000 -

0.2005764 -0.1226404

Proverty.Rate -0.23855481 0.33231611 -0.2005764

1.0000000 0.1225030

Unemployment.Rate -0.13715912 0.11892906 -0.1226404

0.1225030 1.0000000

Since all the correlation coefficient is < 10, hence there

is no multicollinearity.
We plot our model to check for homoscedasticity:
From the plotted data we can say that
homoscedasticity is present. To get rid of it we use the
Box-cox Transformation.

From the
BoxCox transformation ,we get lampda(λ)= 0.7474747
The Multiple Regression Model is given by,
Y^0.7474747= 22.6138677+0.0042366X1+
0.0002684X2 - 0.0414551X3 - 0.0048606X4
As we can see the adjusted R2 is 0.7018. i.e. 70.18%
variation is explained by the model. Now to check for
accuracy we obtain the measurements MAE, MSE,
RMSE, MAPE.
MAE( Mean Absolute Error) : 1.617326
MSE(Mean Squared Error) : 3.76621

RMSE(Root Mean Squared Error) : 1.940672

MAPE(Mean Absolute Percentage Error) : 6.296913

After the boxcox transformation on the context of our

dependent variable range 0 to 100 0.7474747.We consider
that our model is performing reasonably well.
Model Conclusion:
Initially, a multiple linear regression model was fitted:
Y= 64.5218970+ 0.0171450X1 + 0.0010833X2 -
0.1631505X3 - 0.0203367X4
Which resulted in an adjusted R2 0.4154. The
model is indicated a moderate fit, suggesting that the
independent variables(sex ratio, population density,
proverty rate and unemployment rate) explained 41%
of the variability in the literacy rate.
To improve the model’s fit, a Box-cox transformation
was applied, resulting in a lampda( λ)= 0.7474747. This
transformation led to a new regression model is given
by,
Y^0.7474747= 22.6138677+0.0042366X1+
0.0002684X2 - 0.0414551X3 - 0.0048606X4
This transformation significantly enhanced the model’s
performance , increasing the adjusted R 2 to 0.7018,
indicating the transformed model explains 70.18% of
the variability in the transformed dependent variable.
Additionally, the error metrics for the new model were
as follows:
MAE( Mean Absolute Error) : 1.617326
MSE(Mean Squared Error) : 3.76621
RMSE(Root Mean Squared Error) : 1.940672
MAPE(Mean Absolute Percentage Error) : 6.296913
The model shows strong predictive accuracy with a low
MAE of 1.62 and MSE of 3.77, indicating mostly small
errors. The RMSE of 1.94 confirms prediction stability,
while a MAPE of 6.3% reflects good overall precision.
Applying the Box-Cox transformation improved model
fit and reliability by stabilizing variance, enhancing its
ability to analyze factors affecting literacy rates in India

QUALITATIVE ANALYSIS
GOVERNMENT INITIATIVES:
Improving literacy rates in India has been a significant
focus of the government , recognizing that literacy is a
cornerstone for socio-economic development. Here’s
an overview of some of the key initiatives:
 National Policy on Education(NPE), 1986 and
1992
 Sarva Shiksha Abhiyan (SSA),2001
 Mid-Day Meal Scheme(MDMS),1995
 National Literacy Mission (NLM),1988
 Right to Education Act (RTE),2009
 Beti Bachao, Beti Padhao (BBBP),2015
 Samagra Shiksha Abhiyan ,2018
 Digital India Campaign ,2015
  IMPACTS:
 Increased Enrolment and Retention : Programs like
SSA and MDMS have significantly increased school
enrolment and relation rates , particularly in rural areas
and among disadvantaged communities.
 Focus on Girl Education : Intiatives like BBBP and
the RTE ACT have increased the focus on girl child’s
education, helping to bridge the gender gap in literacy.
 Adult Literacy : The NLM and subsequent adult
literacy programs have improved literacy rates among
adults, empowering them to participate more actively
in socio-economic activities.
According to the 2011 Census, India’s literacy rate rose
from 64.83% in 2001 to 74.04%, an increase of 9.21
percentage points. Male literacy increased by 6.88
points to 82.14%, while female literacy saw a larger
rise of 11.79 points to 65.46%. Of the 217.7 million new
literates added during the decade, over 110 million
were females, highlighting significant progress in
female education.
EFFECTIVE LITERACY
COMPARATIVE ANALYSIS
 COMPARISON WITH NEIGHBOURING
COUNTRIES:
Comparing Indian literacy with neighbouring Countries
like Bangladesh , Bhutan, Chaina, Sri Lanka and Nepal
reveals varied strategies and outcome. In 2021, the
literacy rate in India was reported to be 77.7%
according to the National Statistical Office (NSO). Here
is a comparison of literacy rates in 2021 for India and
its neighbouring countries.

Countr Banglad Bhuta China India Myanm Nepal Pakist Sri

y esh n ar an Lank
a
Litea 74.90% 66.60 96.80 77.70 76.40% 67.90 58.90% 92.30
racy % % % % %
Rate
 Neighboring countries have implemented various
education initiatives: Bangladesh emphasizes
universal primary education; Bhutan focuses on free
primary and adult literacy programs; China invests
heavily in compulsory and rural education; Myanmar
and Nepal prioritize access and non-formal
education; Pakistan faces challenges but runs
literacy programs; Sri Lanka’s free education policy
has driven high literacy. India’s literacy rate
surpasses several neighbors but remains below Sri
Lanka and China, highlighting progress yet the need
for ongoing efforts to address regional and gender
disparities.

CONCLUSION
This project analyzed literacy rates across Indian
states, examining the impact of socio-economic factors
like sex ratio, population density, poverty, and
unemployment. It involved data visualization, multiple
linear regression modeling—with a Box-Cox
transformation improving explanatory power to 70%—
and evaluation of government literacy initiatives.
India’s literacy rate (77.7%) surpasses several
neighbors but remains below China and Sri Lanka. The
study highlights significant progress but also persistent
regional and gender gaps, emphasizing the need for
continued policy focus to further improve literacy and
achieve equitable educational outcomes.