TIBCO Spotfire® is a comprehensive analytics solution in the market and makes it fast and easy for a developer to visualize discoveries from the data source. Furthermore, report consumers can dig more value from their organization’s data asset.
This passage is going to deliver an introduction to Spotfire Analytics. The first section would render some basics about the software itself. Then two bar chart implementations will show how to use Spotfire to compose a valuable story practically in part two. After that, several suggestions on report-producing in the last section.
First, Spotfire needs data to launch a new project. Generally, there are five methods to retrieve the data:
- Open data files directly
- Data Connections
- Add tables from the database
- Information Links
- Copy and Paste
To the perspective of how data kept, there are two types: in-memory one and the external one.
- data from files
- Information Link
- Data Connection
- from the copy board.
What the in-memory means is that the data imported is embedded in an analytic hence no matter working offline or not logged in, the data is eternally inside the report.
- Data Connection
- Information Link
The external type denotes that the analytic and the data are two distinct entities.
If the developer persists in using data externally, others who do not have the proper permission of the relating database or Information Link may not be able to review the intact report because of the missing data.
When taking the Information Link or Database approach to supply data, the data consistency issue comes up — when and how to update the data of the report? The solution is — Spotfire does not update the date until the user/developer click s the refresh button.
Visualization Basics–Bar Chart Approach
Visualization enables end-users to inspect data more straightforward and clear, thus make more dependable decisions.
The two primary aims to depict a bar chart is to compare aggregate measures in categorical data and to examine data distribution using continuous (binned) data.
The implementation of a bar chart concerns multiple features and detailed configurations of Spotfire, like most of the other visualizations. Therefore, two instances below reveal how to compose meaningful bar charts.
About the Data Set
Extraction was done by Barry Becker from the 1994 Census database. A set of reasonably clean records was extracted using the following conditions:
((AAGE>16) && (AGI>100) && (AFNLWGT>1)&& (HRSWK>0))
The original prediction task is to determine whether a person makes over 50K a year. Notwithstanding, we exploit the data to obtain more impressive facts about the US people in 1994.
The raw data is a CSV file, so we import it directly from local storage.
Click the third button of the left sidebar.
Edit the X-axis and Y-axis property, i.e., what kind of data we would like to present on this pole. As shown in the picture, the chosen aggregate measure is
Avg(hours-per-week), which means the average working hours per week. Moreover, the X-axis is five different races formed as the classification.
All average working hours per week are between 35 to 40, and the Y-axis is split by 5. If we would like to take a more in-depth observation, a slider may be the right choice.
Besides, labeling the marked part, exposing the tooltip while hanging the mouse over the view, and displaying gridlines on the horizontal axis can assist us in identifying detailed information.
Nevertheless, this diagram does not take advantage of other attributes from the data source, like the sex and the working class. Further, we can split the original X-axis property into two genders, for instance, white male and white female. Moreover, clicking one bar to exhibit the percentage of the work type is an essential function. Here comes a new visualization–pie chart– employed to manifest percentage, we can generate one straight with a right-click.
The action of clicking, selecting an area by drag a square space is named as marked in Spotfire. Once data are marked, it has the possibility that other visualizations change due to original data were substituted by marked one. Click anywhere spare can remove the mark.
It needs to be mentioned that, in this situation, among visualizations, the marking color denotes selected data should be identical.
If reviewed the data and the X-axis accurately, it appears that the other-race bar is missing, and the average working hours for each race altered. It’s not because of the loss of data but another critical feature of Spotfire — Filter. The primary purpose of the filter is to exclude specific data from the current page as if it doesn’t exist. It’s much of help when we would like to clean the undesired data or select a subset of an attribute to procure more enhanced results.
We can find the Filter panel by clicking the View tab, then click the Filter.
As presented from images below, these restrictions narrow down data applied to visualizations:
- Young adulthood (18~40) and Middle adulthood (40~65)
- The unknown, unpaid, unemployed work class is excluded.
- Unknown or mixed race is excluded.
Likewise, we can create another bar chart based on the average capital gain (aggregate measure) and continuous binned age. Yet, the goal of this kind chart is to describe the distribution of data.
What’s more, we can configure the delimiter of each bin by editing the expression in the aggregation function.
When you achieve access to the data source, the first thing to do is not to assemble it into a project but to launch feasibility research:
- Does the data come from a dependable source?
- Has it been processed with a standard ETL procedure?
- Is it well-documented that can acquaint the developer and user what each attribute means?
In a nutshell, the validated data lay a cornerstone on essential reports and visualizations’ details matter.