Semester Project Report
Probability and Statistics
Submitted By: 2021-MC-20
Submitted to: Dr, Abbas
DEPARTMENT OF MECHATRONICS &
CONTROL ENGINEERING
29 May, 2024
Part B – Create a Regression Model using Data Collected from the Real World
a . Explain what bivariate data you wish to collect and explain what is the population
of your study.
I wish to collect bivariate data on the average monthly temperature and the electricity consumption for
residential households. The population of the study will be the monthly temperature and electricity
consumption data.
b. Describe what type of relationship you expect to find between the two variables.
I expect to find a relationship between the average monthly temperature (independent variable )
and the monthly electricity consumption (dependent variable). The type of relationship that I expect
to find is likely to be particularly a quadratic relationship. This is because electricity consumption can
be higher in both very hot and very cold months due to increased use of heating and cooling systems,
leading to a U-shaped curve.
c. Give a reason why you believe the type of mathematical model you selected
is most appropriate for this situation.
The quadratic model is appropriate because heating and cooling demands usually increase non-
linearly with temperature. In moderate temperatures, electricity consumption tends to be lower, while
it rises in both extremely high and extremely low temperatures due to the use of air conditioning and
heating systems, respectively. This results in a parabolic pattern when plotting electricity
consumption against temperature.
d. What type of correlation do you expect to find?
I expect to find a strong correlation between temperature and electricity consumption. This strong
correlation is anticipated because temperature is a primary driver of electricity use in residential
settings, where heating and cooling are major components of energy consumption. The nature of the
correlation is likely to be non-linear , with increased consumption at both ends of the temperature
spectrum.
Example Data Collection
Month Average Temperature (°C) Electricity Consumption (kWh)
January 5 800
February 6 780
March 10 700
April 15 650
May 20 600
June 25 900
July 30 1100
August 28 1050
September 24 850
October 18 700
November 12 750
December 7 810
January 4 820
February 8 770
March 11 690
April 16 640
May 22 620
June 26 920
July 31 1150
August 29 1070
September 23 830
October 19 710
November 13 760
December 8 800
January 3 830
February 9 760
March 12 680
April 17 630
May 21 610
June 27 930
July 32 1200
Final Questions - Answer the following questions in
your report.
4. Based on your scatter diagram, is there a relationship between the x and y
variables? If yes, describe the relationship. Do you believe there may be
association, causation or neither? If you believe a change in one variable causes a
change in the other, then say which variable you believe is the dependent and why.
Yes, there is a clear relationship between temperature and electricity consumption. The relationship
is quadratic, showing that electricity consumption increases in both high and low temperatures. There
is likely a causal relationship because temperature directly influences the need for heating and
cooling, which drives electricity consumption. Therefore, electricity consumption (y) is the
dependent variable and temperature (x) is the independent variable.
5. Describe any correlation. Is it positive or negative? Strong or weak or
moderate? Explain why you think your correlation ended up being what it was.
The correlation is strong and quadratic. It is positive for both low and high temperatures but
decreases in moderate temperatures. This is because extreme temperatures lead to higher use of
heating and cooling systems, while moderate temperatures result in lower usage.
6. What do each of the constants from your mathematical model mean specifically
in relation to the variables you are analyzing? Include units whenever appropriate.
In the quadratic model \( y = ax^2 + bx + c \):
- \( a \) represents the rate at which electricity consumption increases as temperature deviates
from a moderate range.
- \( b \) represents the linear change in electricity consumption with temperature.
- \( c \) represents the baseline electricity consumption when the temperature is at the
midpoint of the data range.
These constants are in units of kWh/°C², kWh/°C, and kWh, respectively.
7. Use your model to make a prediction. Pick a value of the horizontal axis (x)
variable and use your formula to predict an output (y) value. What is the meaning of
your prediction? Is it extrapolation or interpolation and why?
Using the model \( y = ax^2 + bx + c \), let's predict electricity consumption for a temperature of
22°C.
If the model is \( y = 2x^2 + 3x + 600 \) (hypothetical coefficients):
\( y = 2(22)^2 + 3(22) + 600 \)
\( y = 2(484) + 66 + 600 \)
\( y = 968 + 66 + 600 \)
\( y = 1634 \) kWh
This is interpolation because 22°C falls within the range of the collected data.
8. Are the results of the regression what you expected? If not, give a
possible explanation for why it may be different from what you expected.
The regression results were as expected, showing a quadratic relationship. If results were different, it
could be due to data errors, an insufficient range of temperatures, or other unaccounted factors
affecting electricity consumption.
9. What would you do to improve the results of your regression if you were
to continue this investigation?
To improve the results, I would:
Collect more extensive data over a longer period.
Include more variables such as household income, number of occupants, and type of
heating/cooling systems.
Perform regional analysis to account for different climate zones.
10. Give a conclusion of your findings from the report for both parts A and B.
In Part A, the focus was on understanding the relationship between temperature and electricity
consumption, hypothesizing a quadratic relationship. In Part B, data collection and analysis confirmed
this hypothesis. The quadratic model effectively described the relationship, showing increased
electricity consumption during extreme temperatures. Future studies could enhance accuracy by
incorporating more data and additional influencing factors. Overall, the findings highlight the
significant impact of temperature on residential electricity consumption, supporting the need for
targeted energy efficiency measures in extreme weather.
