Evaluating forecasting-based analytics

Summary statistics

Figure 1: Statistics
Source - Created by self

Figure 2: Missing values

Source - Created by self

The summary statistics for each feature are as follows:

Name:
Min: 0.630486
Max: 247.91
Mode: 26.71
Median: 116.358
Dispersion: 29.84
Mean: 109.65
B:
Min: 107.299
Max: 19.8325
Mode: 22.7253
Median: 57.133
Dispersion: 0.439964
Mean: 63.0475
INDPRO:
Min: 0.07
Max: 19.1
Mode: 4.89119
Median: 0.737998
Dispersion: 4.68
Mean: 0.09
FedFundsRate:
Min: 5.7
Max: 54
Mode: 5.95952
Median: 10.8
Dispersion: 0.0264075
Unemployment Date:
Min: 1955-01-01
Max: 2017-01-12
Mode: 1988-01-06 12:00-00
Median: 1955-01-01
Dispersion: 62 years
The absence of missing values practically indicates a robust dataset, while the provided
summary statistics specifically offer a comprehensive view of the distribution of each
variable.
 Figure 3: Orange data mining map
Source - Created by self

ARIMA model forecast

a. The ARIMA model practically predicts that the unemployment rate specifically for
the upcoming quarter will particularly be approximately 4.10938%. This forecast is
specifically accompanied by a 95% confidence interval, therefore indicating a range
of potential outcomes. The lower bound of this interval is specifically estimated at
3.74398%, while the upper bound is 4.75655%. These confidence intervals
particularly provide a measure of uncertainty around the forecast, thus suggesting that
the actual unemployment rate is likely to fall within this specified range. It is
specifically crucial to consider these intervals when specifically interpreting the
forecast, as they offer valuable insights particularly into the reliability of the model's
prediction and also the potential variability in the actual outcome (Khan &
Alghulaiakh 2020).
b.
 Figure 4: Unemployment rate forecast for ARIMA model
Source - Created by self

Figure 5: Unemployment rate 95% CI low for ARIMA model

Source - Created by self
 Figure 6: Unemployment rate 95% CI high for ARIMA model
Source - Created by self

VAR model forecast

a. The VAR model predicts that the unemployment rate practically for the upcoming
quarter will specifically be approximately 4.09572%. This forecast particularly comes
with a 95% confidence interval, therefore providing a range of potential values. The
lower bound of this interval is particularly estimated at 3.47339%, while the upper
bound is 4.71804%. These confidence intervals specifically serve as a measure of
uncertainty around the forecast, therefore indicating that the actual unemployment
rate is particularly likely to fall within this specified range. It is practically crucial to
take these intervals into account when interpreting the forecast, as they particularly
offer valuable insights into the reliability of the model's prediction and also the
potential variability in the actual outcome (Drachal 2021).
b.
 Figure 7: Unemployment rate forecast for VAR model
Source - Created by self

Figure 8: Unemployment rate 95% CI low for VAR model

Source - Created by self
 Figure 9: Unemployment rate 95% CI high for VAR model
Source - Created by self

Fed Funds rate and the unemployment rate affect

In Genovia, the relationship particularly between the Fed Funds rate and also the
unemployment rate is particularly a critical aspect of its economic dynamics. When the
Federal Reserve practically adjusts the Fed Funds rate, it directly specifically influences
borrowing costs practically for financial institutions. A lower rate stimulates economic
activity specifically by making borrowing cheaper, therefore encouraging businesses to invest
and also consumers to spend. This can specifically lead to increased demand for goods and
also services, subsequently driving economic growth and even potentially reducing
unemployment.
Conversely, if the Fed raises the Funds rate, therefore borrowing becomes more expensive,
which can particularly slow down economic activity. This can specifically have a dampening
effect on job creation and also potentially lead to an uptick in the unemployment rate
particularly as businesses may become more cautious about hiring.
Conversely, the unemployment rate can also specifically influence the Fed's decisions on
interest rates. A high unemployment rate practically suggests an underutilised labour force,
which can predominantly lead to reduced consumer spending and also economic stagnation.
To counter this, the Federal Reserve might particularly lower the Funds rate to specifically
spur economic activity and also job creation.
In Genovia, the interplay particularly between these two variables is carefully monitored and
even managed to maintain a healthy balance particularly between economic growth and also
employment stability.

Figure 10: Tableau variables graph visualisation

Source - Created by self

Figure 11: Tableau variables graph visualisation

Source - Created by self
 Figure 12: Tableau variables graph visualisation
Source - Created by self

Figure 13: Tableau variables graph visualisation

Source - Created by self
Forecast by Tableau
The Tableau forecast for particularly the unemployment rate is 5%. This prediction represents
the expected unemployment rate practically for a specific period, specifically based on the
data and also modelling techniques specifically utilised by Tableau. It is important to
practically note that forecasted values are particularly subject to various factors and
uncertainties, therefore including economic conditions, policy changes, and also unforeseen
events, which can specifically influence the actual unemployment rate. Therefore, while
Tableau's forecast practically provides valuable insights, which users should consider
alongside other economic indicators and also data sources to specifically gain a
comprehensive understanding primarily of the employment landscape.

Figure 14: Tableau Predictive Median model graph visualisation

Source - Created by self

Model used in Tableau

Tableau utilised the Predictive Median model particularly for its forecast. This model
particularly aims to estimate future values specifically by identifying the median point from
historical data. It doesn't practically rely on complex statistical techniques or even algorithms,
making it straightforward and also easy to implement. However, it may not practically
capture nuanced patterns or even outliers that more sophisticated models can specifically
discern. In this approach, the central tendency of the historical data particularly serves as the
basis for the prediction, therefore providing a simplified yet useful method particularly for
generating forecasts. Users should be specifically aware that while this model offers
simplicity and also ease of interpretation, it may not practically be as robust in capturing
complex trends or even sudden deviations in the data.
Best forecasting model
 Figure 15: RMSE and MAPE of each model
Source - Created by self

Among the three models specifically evaluated, the ARIMA model particularly provides the
most precise forecast predominantly for the unemployment rate. It predicts an unemployment
rate of approximately 4.10938% specifically for the upcoming quarter, which is supported by
a 95% confidence interval ranging from 3.74398% to 4.75655%. This particularly indicates a
relatively narrow range of potential outcomes, therefore suggesting a higher level of
confidence in the forecast. In contrast, the VAR model specifically predicts a slightly lower
unemployment rate of 4.09572%, practically with a broader confidence interval of 3.47339%
to 4.71804%. Tableau's forecast, practically while straightforward, thus estimates a higher
unemployment rate of 5%, thereby without providing a confidence interval. Overall,
practically based on the provided metrics, the ARIMA model particularly demonstrates
superior forecasting accuracy particularly in this scenario.
Summary
The report specifically provides comprehensive insights particularly into the dataset,
including summary statistics specifically for key variables. It practically highlights the
absence of missing values, therefore indicating a robust dataset. The ARIMA model
specifically forecasts an upcoming unemployment rate of approximately 4.10938%,
practically with a 95% confidence interval. The VAR model particularly predicts a slightly
lower rate of 4.09572%, also specifically with a confidence interval. Tableau's predominantly
forecast estimates a rate of 5%. Among these models, the ARIMA model specifically
demonstrates the highest precision, therefore offering a narrower confidence interval, thus
making it the preferred choice for forecasting particularly the unemployment rate.
References
Drachal, K 2021, ‘Forecasting crude oil real prices with averaging time-varying VAR
models’, Resources Policy, 74, p.102244. retrieved 16 September, doi:
https://doi.org/10.1016/j.resourpol.2021.102244
Khan, S, & Alghulaiakh, H 2020, ‘ARIMA model for accurate time series stocks
forecasting’, International Journal of Advanced Computer Science and Applications, vol. 11,
no. 7. retrieved 16 September,
https://pdfs.semanticscholar.org/0f86/bc34afa9f222873cc42c4516746a3911342d.pdf