[JAVA] Let's try Zoomdata (extra edition)

We have introduced the introduction and use of ** Zoomdata ** in three parts in the past, but this time we will introduce ** Fusion ** as a convenient function of ** Zoomdata ** as an extra edition. I will introduce you. I think the word ** microquery ** came up in the previous introduction, but another aspect of this feature is ** when you need it, as much data as you need. Taking advantage of the characteristic of going to see the source **, a virtual join of data with a common key that exists between different data sources ** (because it does not actually rewrite the actual data as a result) and its ** Realization of high-speed visualization analysis **. Also, by making good use of this function, data that is well set and operated with a common key among data sources of ** different big data scattered in the right place ** (in this case ** data lake) Beyond the scale of ** ... ** It may be a scale that must be called Data Ocean ** ...) can now be handled transparently as a single source. It is possible to improve the operational efficiency of ** big data ** and to acquire added value newly created from it.

(1) Preparation

As for the data used in this experiment, I had an image of the output of ** IoT ** type sensor data, so I will divide that data into several parts. (Since it is a confirmation of the basic mechanism, please create test data by referring to the following procedure)

As for the number of data, 6360 data are sampled from each of the 4 sensors, and for convenience, a sequential sample ** ID ** is added in chronological order. By the way, the chart that reads this data as it is and visualizes it

So, in the end, if you fuse ** 4 sensor data and put them together in the same display chart, it will be a goal. Now, let's divide the data first. This time, it was divided into 3 parts: Unit-1, Unit-2, Unit-3 & 4.

(2) About ** Fusion ** settings

First, enter ** Zoomdata ** with ** admin ** to configure the data source. (The procedure is the same as before, so it is omitted)

When you're done, select ** Fusion **. (It is on the far left of the ** Source ** list)

Select ** Next ** and select the data for ** Fusion **.

Select ** Next ** to start the specific settings.

Since the data ** ID ** will be used as common information this time, the ** ID ** attribute will be extracted from each original data and registered by dragging and dropping.

Now that the basic structure is automatically defined, proceed in the same way.

Select ** Next ** when you have finished configuring the three data sources.

The table generated according to the instructions will be displayed, and you will be able to make more detailed settings, but this time the purpose is to introduce the flow of operations quickly, so select ** Save & Next ** as it is ..

The ** Time Bar ** setting screen will be displayed. For convenience, set the synchronization with the time data of the first data source and select ** Finish **. When the work is completed successfully, the top page of the familiar data source will be displayed.

(3) Creation and verification of charts

Next, let's check if the same output as the chart at the beginning can be obtained. Select the leftmost icon from the icons on the left side, and make sure that ** Fusion Data Source ** is in the list of ** Data Source ** and select it.

The available charts will appear, select ** Line Trend: Multiple Metrics **.

The first graph will appear, so increase the visualization target.

Change the time unit at the bottom of the chart from ** DAY ** to ** MINUTE **.

The visualization chart is now the same. I will change it from ** MINUTE ** to ** HOUR ** this time to check if it is linked.

The ** Fusion ** data is completely under the control of ** Zoomdata **, so you can see that redrawing under different conditions can be done smoothly.

(4) Summary of this time

This time, it's a trial extra edition, so I introduced the basic procedure for ** Fusion ** in a simple form. As a point (1) Can be used between different data sources as well (2) With the ** microquery ** effect, high efficiency and performance can be maintained even between ** big data **. (3) The ** Fusion ** structure can be flexibly changed even during operation. And so on.

In the case of a mechanism that assumes the handling of ** big data ** via the conventional ** full query **, it is very difficult to constantly realize efficient and practical handling. On the contrary, I think the reality is that we have devised a design and operation that does not generate such demands. ** "There are new strategies, added value, and new clues to solve problems in the process of creating the final form of data" **, a big data-oriented idea **, and entered the practical domain * * When considering efficient cooperation with machine learning and AI **, how to approach areas such as data collection, storage, visualization, and analysis is important for data handling in future corporate and group activities. You may have come to a stage that will be a key point.

** It is very difficult to handle data that does not exist in the past **, but ** Capture as much data as possible in real time, visualize and analyze the process at high speed, and flexibly verify temporary construction on-stream. The realization of a new mechanism ** that can support on-site response, quick decision making, and action by making corrections and changes to ** is the new future of ** big data **.

** The fact is that it's not happening in the conference room, it's happening in the field ... ** The one who controls this reality can win the next market domination and efficient problem solving. It may be.

Next time, I would like to introduce the cooperation with big data.