Data visualization in OCDSearch
Initially, we planned to build OCDSearch as a search engine only. But after first weeks of alpha testing, we realized that it could be much more useful if we add some basic data visualization tool.
For this, we used Kibana, an open-source data visualization tool. It could be a bit less flexible or powerful, comparing to market leaders from QlickTech, Oracle or Microsoft, but it has several advantages:
- It is free;
- It seems to be fast enough (at least, at first glance);
- It uses data, indexed by the search engine. In other words, this is full-text search and business intelligence tool in one box.
Like any other business intelligence tool, it is interactive. In other words, when you select any item of the dashboard, it automatically recalculates other elements. Let’s see it on a practical example. When you open Colombia’s page, it will look like this:
Then you may decide to select only open tenders and click on the green section of the pie chart in the upper left corner. When you do this, your dashboard will look like this:
Structure of the statuses on the pie in the upper right corner changed. So did distribution of tenders between buyers on the bar chart in the center. You can proceed more and more applying additional filters and you diagrams will change.
You would say “nothing new,” and you will be right. Many applications do the same. Okay, but what do you think about this feature?
Full-text search + business intelligence: all-in-one
Most of the BI tools which we know work pretty straightforward: there are dimensions, which can be used for analysis of the data, and which must be pre-defined. You can aggregate transactions by these dimensions and play with it. Often, this is it.
In our case, there is a bit more flexibility. Look at the search field at the top of the page. You can add any word or even a clause and see what will happen. Let’s check the word ‘construcción’. Here are tenders, containing this word:
Let’s try to find something more specific. For example, trademark ‘mitsubishi.’ Here we are:
And just a few words about ways to customize your dashboards. There are 16 different visualizations, including gauges, time series, maps, etc. Some of them are easy. Others require some knowledge. But all of them can be set up without programming. For example, adding two metrics to the dashboard takes less than two minutes. Now the dashboard looks like this:
Looks pretty easy, yet useful.
We are still testing this tool. As of today, we are more than excited. Let’s see what will we say in some time.