Hi guys! Julie here again! Today we’re continuing our exploration of indico’s APIs. As a quick reminder, I’m the non-programmer Ops Manager, serving as your guide to these newfangled machine learning tools that can help you improve your company and your work in many ways, such as developing better brand strategy or creating new business opportunities. Also: to provide more cat GIFs because as we all know, cat GIFs are truly what powers the tech industry (and we’re all about a better tech industry here at indico).

Before we dive in, let’s remember what this series is all about:

  1. Machine Learning is for everyone, and with indico’s APIs you don’t need a PhD to put it to work for you.
  2. I’m going to show you how I (a non-technical person like yourself) use these APIs to create work faster and make it better.
  3. #Cats

Today we’re going to look at Text Tags. This is a great way to get a general idea of what a body of text is really all about.

I have two cats, and my philosophy is that cats should rule the world. I mean, they already rule the internet, right? They definitely rule my house. Here’s my regal cat, Jack, in this very fine piece of art I made with the FORK app.

machine learning cat

So let’s see my life in action. Let’s come up with a sentence. How about…“Why cats should rule the world”?Screen Shot 2015-10-06 at 3.23.57 PM

Here you can see that the API has determined the top three topics that are most relevant to my sentence, with “pets” and “philosophy” outweighing “comics”. Pretty darn good. It’s great for getting a quick look at large volumes of punchy text — like the kind you’d find on Twitter or Facebook. For example, you could use this API to figure out what your customers’ interests are by analyzing the topics they discuss the most on social media. From there, you could improve your products to better suit their needs, or fine tune your marketing and advertising strategy to resonate with their interests. You can see how this might work by playing with our Twitter Thumbprint tool.

What about longer pieces of text? I follow a ton of blogs, and many (but not all) of them include some kind of tagging system where I can click on the tags to view more articles relating to that topic. But I don’t want to have to spend time searching through each individual site to find all the new articles about pets, and then manually compile them for later reading. And what about the ones without tags? I’d have to make a guess from the title or read a synopsis to figure out whether it falls under the topic I’m looking for.

Instead, the Text Tags API reads through all ten of the different blogs I follow, and then generates a daily, customized feed of articles that interest me. And voilà, there’s my rush hour reading all lined up. Time to move on to the next tweet, blog post…or cat GIF. In fact, hasn’t it been at least two paragraphs since we’ve seen a cat?

via GIPHY

Ruling the world is hard work.

So we’ve seen how the Text Tags API works on a short piece of text. Let’s try something a little longer. Here we have our previous blog post (in which we introduced you to our super cool Twitter Thumbprint and Sentiment Analysis tools). Let’s run the article through our Text Tags tool and see what we get!

text tags

You can see that it took my 1,000 word cat GIF/machine learning opus and boiled it down to the gist of what it is about: startups, programming, and math. Pretty good analysis of the high level themes in that blog post. But seriously, that’s balla. I didn’t do anything but copy and paste.

I know you’re all itching to put this into practice, or get the programmers on your team to integrate this into your work. I can hear you saying right now, “Julie this is so great, I love this API and I think it would make my life so much easier, and also I love cats and I want to meet Jack and Ben! And I also think we should meet up for a martini and talk about my cats too. So yes! Let’s meet! But how am I going to convince my programming team to use this? They’re very particular.”

The answer is simple. indico’s API’s only take a few of lines to code deploy, and we have super easy access in lots of popular programming languages. It’s barely an integration. Don’t believe me? Just show them this demo and then pick their jaws up off of the floor.

via GIPHY

This is all it takes to put the power of machine learning into your software.

So, if you’ve got a whole bunch of people talking about a lot of different things, there’s no need to painstakingly read and sort through all of that information yourself. Just use indico’s Text Tags API.

If you want to make this tool even more powerful, you can use it with our Keywords and Named Entity Recognition APIs. #2isBetterThan1! More on those later in our series.

indico’s APIs are your one stop shop for making management of your everyday electronic footprint that much easier for you. Do you have more questions and want to ask them in a safe place with no judgment? Don’t go to your information dealer on the corner. Hit me up at julie@indico.io. Follow me on Twitter at @southendjulie and @indicodata while you’re at it.

See you next week!

Suggested Posts

Is Artificial Intelligence Within Reach?

Using Passage for Text Analysis with Recurrent Neural Networks at Next.ML Cambridge

SAS: The Unreasonable Benefits of Deep Learning