Reporting with Hadoop - Big Data Made Easy: A Working Guide to the Complete Hadoop Toolset - page 332

Many of the reports shown in Figure 11-10 are built into Hunk, but I created the fuel types, sales volumes, and
manufacturers reports. When I save a report, I can choose to add it to a dashboard that I can create at the same time.
For example, in Figure 11-11 , I create a dashboard called “Vehicle” to contain the three reports I've just mentioned.
Figure 11-11. Hunk dashboard for reports created
To open the dashboard, I choose the Dashboards menu option, then select the Vehicle dashboard. The display
is populated from the current data, giving me, on a single page, an up-to-date view of the state of the data. A simple
example of what can quickly be created to represent HDFS-based data, Figure 11-11 shows two pie charts that display
top manufacturers and fuel types. The bottom bar graph displays sales volume by year. Once the report entries are
on the dashboard, I can edit them to add titles, or drag and drop them into the best positions to give an at-a-glance
impression of the data status.
I can also create complex report expressions to construct multi-column reports. For example, the following
report expression creates a table that contains the minimum CO 2 emissions by manufacturer and model and was
supplied by Ledion Bitincka, an architect at Splunk, in response to a question I submitted to the Splunk forum):
index=cdh5_vindex | stats min(c02_g_km) AS minco2 BY manufacturer, model | sort 20 minco2. Thus, the
original search expression index=cdh5_vindex is passed to a stats function that finds the minimum CO 2 emissions
value from the column c02_g_km in grams per kilometer and saves it to a new column name called minco2 . The results
placed in this column are grouped by manufacturer and model, just as you would group non-aggregated columns in
SQL. Finally, the data is sorted and the top 20 rows are displayed, creating the report shown in Figure 11-12 .