KNIME: A Living Software Platform
We started using KNIME simply as a tool to achieve our goal, but today it is more than that.
If at first, as we noted in a previous post, “KNIME, a miner for the mountain of data”, we talked about how easy this software makes things for business users, today we can say without a doubt that KNIME brings a great deal of value to any type of user, as it is becoming increasingly easy and straightforward to use.
KNIME Analytics Platform
Today we are going to evaluate and analyse the new options this software includes in its latest update.
First, we will highlight the addition of new charts to this tool. While it is true that it is not a powerful data visualisation tool like Qlik or Tableau, we are seeing how visualisations are gradually gaining importance within KNIME. Three new ways of visualising the data have been added: “Network Viewer”, which represents data as a graph; “Sunburst Chart”, a chart divided into slices and sub-slices that categorises the data within each one; and “Streamgraph”, which represents data in a flowing stream form. Below, we show a data representation using the “Sunburst Chart” model.
In addition to viewing the chart, we can select the regions of interest and see the data corresponding to each section, as shown in the following image.
Applied to the business world, if we think of a warehouse with a large number of SKUs, we could represent the different product families in the smallest circle. In the next circle out, we could represent the groups within each family and, finally, in a third circle we would represent each of the items belonging to each group. In this way, we could quickly and effectively identify visually which group and which family an item belongs to. We could also see whether two items belong to the same group or the same family.
Version control
Another improvement KNIME brings in this update is the ability to compare the changes made in a workflow across versions and see easily, quickly, and conveniently what has changed over time. In this way, if the result has changed and is not the expected one, we can roll back, knowing at which step we made the mistake.
In KNIME, it is easy to use the same metanode multiple times, and until now we would copy the metanode and paste it wherever appropriate. With the new update, we can save these metanodes as if they were a Word template. We save them in the Templates section and, whenever we want to use them, we simply drag them in, just as we have done up to now with nodes. We drag them into the workflow and connect them to the required nodes. This new update will save us time and, consequently, money.
New nodes
A new node called “Call Local Workflow” has been added, used to make a call from one workflow to another that exists locally. Until now, the only way to run two workflows where one depended on the other was to run the first one to prepare the data for the second and, once it finished, call the second workflow. This avoids the arduous process of matching all the nodes and finding a way to connect them.
Finally, we mention speech recognition and analysis: in this update, KNIME incorporates a speech processor capable of listening to audio files from within KNIME, extracting acoustic features, and performing speech recognition.
In this way, without having extensive IT knowledge, we can use recordings as if they were a simple CSV and, from there, carry out our own analysis of the results and draw conclusions. This new addition could even help us create a small voice-activated function system.
With this new update, KNIME remains at the forefront as a Big Data tool and continues to reach an increasingly broader audience, because it is able to make the most complex and cutting-edge technology—Big Data—simple and easy for any profile within the company.
For these and many other reasons not evaluated in this post, at LISDATA we continue to rely on tools as versatile, useful, and constantly evolving as KNIME.