C207 Data-Driven Decision Making: Bitcoin Price Prediction Report

This report presents a data-driven analysis of Bitcoin price forecasting, focusing on the relationship between Bitcoin's daily price, volume, and market capitalization. The study utilizes historical data from January 24th, 2019, to March 23rd, 2019, employing multiple regression analysis to develop a predictive model. The analysis reveals a strong positive correlation between Bitcoin price and both volume and market capitalization. The model, derived from the regression analysis, is formulated as: Daily Price = -2.76E-09 (Volume) + 5.72E-08 (Market Cap), which demonstrates the potential for forecasting Bitcoin price changes based on market data. The report highlights the model's high explanatory power (R-squared = 0.999888701) and its applicability in informing investment decisions, allowing investors to predict potential price fluctuations and adjust their strategies accordingly. However, the report acknowledges limitations, such as not considering external factors like interest rates and political stability.

Task 2
Data-Driven Decision Amking-C207
Student name:
ID number:
Date:

Summary of the real-world business situation
The emergence of numerous cryptocurrencies has attracted individual and company investors.
Bitcoin is one of the most famous and valuable Cryptocurrency. Investors and business analysist
are trying to determine the best investment decisions that can result into high returns.
Determining the expected price change (rise or fall) is a major factor in Cryptocurrency
investment (Borgonovo, et al., 2018). The change in prices can be predicted/forecasted using the
historical data. The other factor that can be determined is the relationship between price and
other factors such as volume and market cap. Therefore, the research question that we will seek
to answer is: Can we forecast the price of Bitcoin using the market capitalization and volume?
Description of the Data
The data that that has been used for analysis consists of the information about the historical
prices of Bitcoin Cryptocurrency. The attributes of the data include the average daily price of
Bitcoin, the daily volume of Bitcoin and the daily market capitalization. All the attributes are
numerical in nature and measured in terms of United States dollars. The data has 59 data points
(entries) representing two months between: Jan 24th 2019, to March 23rd 2019.
Graphical display
The diagram below represents a multiple line graph of the three attributes: The price, volume and
market capitalization of Bitcoin for the two periods. The multiple line graph shows the daily
movement of the three attributes as well as the trend over the two months. From the graph, it is
clear that the price of Bitcoin was relatively constant over the two months period. On the other
hand the volume and market capitalization of Bitcoin experience a slight but consistent increase
over the two months period.

24/Jan/19
29/Jan/19
3/Feb/19
8/Feb/19
13/Feb/19
18/Feb/19
23/Feb/19
28/Feb/19
5/Mar/19
10/Mar/19
15/Mar/19
20/Mar/19
0.00
10,000,000,000.00
20,000,000,000.00
30,000,000,000.00
40,000,000,000.00
50,000,000,000.00
60,000,000,000.00
70,000,000,000.00
80,000,000,000.00
A Multiple Line Graph of Price, Volume
and Market Cap
Price Linear (Price)
Volume Linear (Volume)
Market Cap Linear (Market Cap)
Date
Value
Descriptive Statistics
The descriptive statistics analysis provides the summary statistics of the attributes. The summary
statistics are shown in the table below. From the table below, it is clear that the average daily
price over the two months period was $3776.48, the average daily volume was $7842095709.51
and the average daily market capitalization was $66350450021.17. The minimum daily price was
$3411.09 while the maximum daily price was $4067.51.
Price Volume Market Cap
Mean 3776.48 7842095709.51 66350450021.17
Standard Error 27.15 248926024.86 495144386.00
Median 3853.37 7826525254.00 67683296223.00
Standard Deviation 208.55 1912037077.39 3803276195.10
Sample Variance 43494.64 3655885785320590000.00 14464909816181600000.00
Kurtosis -1.35 -1.47 -1.26
Skewness -0.29 -0.04 -0.27
Range 656.43 5826249979.00 13142791217.00

Minimum 3411.09 5004962683.00 59578075991.00
Maximum 4067.51 10831212662.00 72720867208.00
Sum 222812.59 462683646861.00 3914676551249.00
Count 59.00 59.00 59.00
Correlation
Correlation analysis outlines the relationship between the attributes. From the results below, the
correlation coefficient between the daily price of Bitcoin and the daily volume of a Bitcoin is
0.869188. The correlation is a strong positive correlation. The correlation implies that an
increase in the volume of a Bitcoin by 1 unit will cause a corresponding increase in the price of a
Bitcoin by 0.869188 units and vice versa (Bolin, 2014).
The correlation coefficient between the daily price of a Bitcoin and the daily market
capitalization is 0.9825. The correlation is a strong positive correlation. The correlation implies
that an increase in the market capitalization by 1 unit will cause a corresponding increase in the
daily price of a Bitcoin by 0.9825 and vice versa (Bolin, 2014).
The correlation coefficient between the daily market capitalization of a Bitcoin and the daily
volume of a Bitcoin is 0.865859. The correlation coefficient signifies a strong positive
correlation. The correlation coefficient implies that an increase in the daily volume of Bitcoin by
1 unit will cause a corresponding increase in the value of market capitalization by 0.8659477
units and vice versa (Bolin, 2014).
Price Volume Market Cap
Price 1
Volume 0.869188 1
Market Cap 0.982553 0.865859477 1

Data analysis
The appropriate analysis technique that has been used is the multiple regression analysis. The
multiple regression analysis has been done to predict the daily price of a Bitcoin using the daily
volume and the daily market capitalization. The output of the multiple regression analysis is
shown below.
The summary output demonstrates the multipole regression coefficient (R) is 0.999944. The
Multiple regression represents the gradient or the rate of change of independent variables
(volume and output) with respect to the dependent variable (the daily price) (Bolin, 2014).
The value of R square is 0.999888701. The value of R square demonstrate that the regression
model explains 99.9888701% of the population. Therefore, it is clear that our model is a perfect
model that can be used for making statistically accurate inferences about the population (Bolin,
2014).
Regression Analysis
SUMMARY OUTPUT
Regression Statistics
Multiple R 0.999944349
R Square 0.999888701
Adjusted R Square 0.982342889
Standard Error 40.59489239
Observations 59
The table below is the output of analysis of variance (ANOVA). Analysis of variance is
statistical tool that investigates whether there is any significant difference in the mean values of
the attributes or variables. Therefore, the output below investigates whether there is any

significant difference in the average daily price, volume and market capitalization of Bitcoin.
The significance value is 1.2E-111 which is less than 0.05 hence it is clear that there was a
significant difference in the average values of the three variables.
ANOVA
df SS MS F Significance
F
Regression 2 843877028.3 4.22E+08 256039.2 1.2E-111
Residual 57 93932.88143 1647.945
Total 59 843970961.2
The coefficients table below outlines the specifications of the model. From the table below, the
model for predicting the daily price of Bitcoin can be derived as follows:
Daily Price = -2.76E-09 (Volume) + 5.72E-08 (Market Cap)
Coeffici
ents
Standard
Error
t Stat P-
value
Lower
95%
Upper
95%
Lower
95.0%
Upper
95.0%
Intercep
t
0
Volume -2.76E-
09
3.46E-09 -
0.798
0.428
1
-9.7E-
09
4.17E-9 -9.7E-09 4.17E-09
Market
Cap
5.72E-
08
4.207E-10 136.0
23
2.58E
-73
5.64E-
08
5.81E-
08
5.64E-08 5.81E-08
The figure below is the output of normal probability resulting from the regression model. From
the figure below, it is clear that the predicted values of daily price are not normally distributed. A
normal distribution displays a bell- shaped curve.

Justification of using Multiple Regression Analysis
Multiple regression analysis is appropriate in this case
because the problem at hand is a prediction problem.
Furthermore, the data set contains more than one
variable and hence a multiple regression analysis is
appropriate (Bolin, 2014).
Discussion of Results
The purpose of the report was to investigate the relationship between three variables of Bitcoin:
The daily price, the daily volume and the daily market capitalization. The existence of the
relationship among the three variables was to be used to further investigate whether a model for
predicting daily price of Bitcoin can be developed.
The results of analysis clearly demonstrates that the three variables are strongly correlated.
Similarly, the results demonstrate that the model for predicting the daily price of a bitcoin given
the daily volume and the daily market cap is given by:
Daily Price = -2.76E-09 (Volume) + 5.72E-08 (Market Cap)
Therefore, the model can be used to solve the real-world challenge of making informed
investment decisions. The model can be used by business analysists and individual investors in
forecasting the possibility of fall or rise in the price of a Bitcoin. Knowing whether the price of a
Bitcoin will fall or rise at any given time can lead to appropriate investment decisions. For
example, when the prices are expected fall, an investor may sell their Bitcoins in order to avoid
0 20 40 60 80 100120
3000
3500
4000
4500
Normal
Probability Plot
Sample Percentile
Price

suffering the associated loss. On the other hand, when the price is expected to rise, the investor
may buy more Bitcoins in order to gain more profits.
The limitation of this data analysis is that it there are other factors determining the price of a
Bitcoin that have not been taken into consideration. For example, the prevailing interest rates and
the political stability of the economy which have been assumed to be the intervening variables.
The recommendation for using the model is that it must be used only with daily data. Therefore,
any user must be cautious not to use any other period of data as that may result into erroneous
results.

References
Bolin, J. H., 2014. Introduction to Mediation, Moderation, and Conditional Process Analysis: A
Regression-Based Approach. Journal of Educational Measurement, 51(3), pp. 2-10.
Borgonovo, et al., 2018. Between Cash, Deposit and Bitcoin: Would We Like a Central Bank
Digital Currency? Money Demand and Experimental Economics. SSRN Electronic Journal, 1(1),
p. 6.