InetSoft Webinar: Tools to Simplify the Building of BI Applications

This is the continuation of the transcript of a Webinar hosted by InetSoft on the topic of "The Movement Towards Using Unstructured Data in Business Intelligence Solutions." The speaker is Mark Flaherty, CMO at InetSoft.

What are some examples of unstructured data BI applications there besides just trying to search and find a document or a report? Are people saying I want to count up the number of emails that reference a private plan, or this kind of thing?

Using tools and processes, this primarily means tools to simplify the building of the BI application, that will make the process easier for these organizations and their users. They need to assess what data types they’re going to integrate. I think there’s still plenty of opportunity to pick and choose from data types that you could more easily support and incorporate.

But ultimately you do want to have a plan in place as to what data types you have to support in the long run. Identifying the systems that the information has to work with, I think that you’re essentially doing that in bringing the information together via data mashup. Their existing capabilities might be inadequate, and I still think that would be a big challenge. And we talked a little bit about how you analyze this unstructured information, so you can give the context to the unstructured data, the number of occurrences and the word counts and things like that.

You can go even deeper, I don’t think it would be your first step necessarily but people do things like sentiment analysis on text. But it’s still pretty early days in that you should at least be aware that companies are starting to incorporate some sentiment analysis in analyzing social media. If you have a brand name and you want to monitor all the tweets about your brand.

#1 Ranking: Read how InetSoft was rated #1 for user adoption in G2's user survey-based index Read More

And at least make some high level analysis of whether those tweets are positive or negative. My opinion is that the analysis is not very accurate. I’ve seen some products try to do this, and they’re pretty primitive. However, as a first measure, it’s more valuable than doing nothing. I think. We’ll see a lot of evolution and improvement. For instance if you see in your automated analysis using one of these tools, and there are smaller vendors that do it as well, but if your automated analysis shows over time pretty consistently 65 or 75 percent positive, some percent neutral and some percent negative and then all of a sudden you notice that the sentiment drops to 65 percent, it’s worth looking at.

While the numbers may not be accurate in and of themselves. The change is probably accurate. Because whatever technique they were using to assess positive or negative, whether they got it a hundred percent correct or not, the drop is probably real. Did it drop 10 percent or did it drop 3 percent? I don’t think they can capture that very accurately yet. But if people are using negative words, and it’s very simplistic now. Most of these analyses simply look for words like: bad, ugly, slow, negative terms like these. You create a library of terms that you consider negative, and if anybody uses that word in the tweet, it’s characterized as negative. Now suppose somebody says ‘that demo was awesome. It was really bad.’ They’re not going to get that one right. But still I think that directionally if you see a drop from 75 percent positive to 65 percent positive; there’s probably something going on. And at least it prompts you to look further.

Deploying these BI applications in the context of collaboration, that is something we’re recommending to any user organizations If you’re delivering your dashboards on a mobile device, the work you do to incorporate unstructured information will be relative easy to also deliver via the mobile device. Enterprises should look at this as part of their overall strategy.

There is a good point that unstructured information will evolve to include really unstructured information like voice or video or audio. But video and audio will probably become more important sources of information. And there is a place where they’re already very important, and that’s in the contact center applications. So to the extent that that market would be more interested in not only analyzing texts and unstructured documents but also analyzing the recordings of their customer support calls and things like that.

There are contact center vendors whose applications are doing a couple of things. Some are merely capturing the recordings. So if you want to go back and hear the recording you have got it available. There are some that are transcribing the calls in an automated way, transfoorming voice to text And it’s like OCR, it’s not going to be a hundred percent accurate, but it gets you some percentage of accuracy. And there are those that are not only capturing the information, transcribing it, but also doing some sort of analysis across that for, again the text analytics. Once it’s been converted to text, you can do the standard text analytics. Again, this is happening in a small minority of the market, and I only mention it id you’re mapping out a strategy, something to consider as part of your strategy. Again, I wouldn’t start there.

Read what InetSoft customers and partners have said about their selection of Style Scope for their solution for dashboard reporting.

My objective in this Webinar was to say that long term, I think you’re going to need a strategy to deal with both structured and non structured information. I don’t know if that’s a year from now, two years from now, five years from now.

Previous: Chat Tools to Track Decision-making