IEEE Power & Energy Magazine - May/June 2018 - 25

Concluding Remarks
2.5
Power (GW)

2.0
1.5
1.0
0.5
0.0
1

18
24
30
Lead Time (h)

Scenarios
Point Forecasts
Observations

figure 9. Probabilistic forecast information conveyed by
ensemble forecasts for the whole wind power generation of
western Denmark.

simple and intuitive way to see how uncertain the forecasts
are, while the latter is provided with enough information to
reconstruct full predictive densities to be used as input to a
wide range of decision and control problems in a stochastic
optimization framework. note that figure 8 does not mean
to show accurate wind power forecasts, so readers may ignore
the fact that many observations are falling outside the 90%
prediction interval.

Ensemble Forecasts
While the visualization in figure 8 is appealing, it is not the
only way to communicate probabilistic forecast information.
instead of the uncertainty of the future, an alternative approach
provides the forecast user with a set of alternative trajectories
in the future. this approach was championed by the meteorological community, which coined the term "ensemble forecast"
for it. in practice, this has translated to a number of high-value
applications, for instance related to trajectories of storms and
cyclones and their potential impact.
for renewable energy generation, this type of representation has attracted increased interest due to the additional
information it conveys, allowing the use of these alternative
futures as input to existing tools for operations and control within a deterministic framework. figure 9 depicts the
ensemble forecasts that are used to convey the probabilistic
forecast information for western Denmark, for a given day
in the past. since they are based on related methods, the
general probabilistic information shown in figures 8 and
9 has similarities, especially in terms of trends and uncertainty levels. However, the ensemble forecasts in figure 9
provide additional information in terms of dependencies
among lead times, which is not conveyed by the river-ofblood fan charts.

may/june 2018

in this article, we have offered a few examples of visualizing
big-energy data. although these examples spread across
distribution (smart-meter data), transmission (pmu data),
and generation (wind-power forecast data) and cover both
pre- and postmodeling stages, we do not attempt to be comprehensive. there are many other insightful plots we did not
present, such as maps for geospatial information (e.g., load
growth and penetration of electric vehicles). some insights
are better presented dynamically via animation rather than
on a static page, such as changes of load and temperature
relationship over time and customer behavior changes due
to the adoption of demand response programs. We hope that
this article can inspire more researchers and practitioners to
create effective plots from energy data.

For Further Reading
m. Belkin and p. niyogi, "laplacian eigenmaps for dimensionality reduction and data representation," Neural Comput., vol. 15, no. 6, pp. 1373-1396, 2003.
r. J. Hyndman and Y. fan, "sample quantiles in statistical packages," Amer. Statist., vol. 50, no. 4, pp. 361-
365, 1996.
X. liu, D. laverty, r. Best, K. li, D. J. morrow, and
s. mcloone, "principal component analysis of wide area
phasor measurements for islanding detection: a geometric
view," IEEE Trans. Power Delivery, vol. 30, no. 2, pp. 976-
985, 2015.
X. liu, J. Ken nedy, D. laver t y, D. mor row, a nd
s. mcloone, "Wide area phase angle measurements for
islanding detection: an adaptive nonlinear approach,"
IEEE Trans. Power Delivery, vol. 31, no. 4, pp. 1901-
1911, 2016.
J. m. mo r a le s , a. conejo, H. madsen, p. pinson,
and m. zugno, Integrating Renewable in Electricity
Markets: Operational Problems. Series in Operational
Research & Management Science. new York: springer
Verlag, 2014.
r. Bessa, c. möhrlen, V. fundel, m. siefert, J. Browell,
s. H. el Gaidi, B.-m. Hodge, u. cali, and G. Kariniotakis,
"towards improved understanding of the applicability of uncertainty forecasts in the electric power industry," Energies,
vol. 10, no. 9, article no. 1402, 2017.

Biographies
Rob J. Hyn d m a n is with monash Business scho ol,
australia.
Xueqin (Amy) Liu is with Queen's university Belfast,
united Kingdom.
Pierre Pinson is with the technical university of Denmark, Denmark.
p&e

ieee power & energy magazine

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - May/June 2018