Five Inspiring Watson Hacks
There is an insane amount of information about IBM Watson on the internet. If you’re a business leader who just wants to know about how Watson can help your business, please stop reading this and go to https://www.ibm.com/watson. If you’re a developer who just wants to know how to get started with the Watson Developer Cloud, again, please stop reading this and go to https://www.ibm.com/watsondevelopercloud. My objective in this article is simply to inspire anyone reading by sharing some of the coolest, out-there, crazy and creative Watson hacks.
Still with me? Okay. Let’s define stuff.
A hack, for the purpose of this blog, is a project that is built during a hackathon. A hackathon, again for the purpose of this blog, is an in-person or online event during which all kinds of developers, designers, business people and others get together, form teams and build prototypes to compete for prizes by winning challenges, which are usually sponsored by various companies.
In my Developer Relations role at IBM Watson I go out looking for interesting hacks that make use of the different services freely* available as part of the Watson Developer Cloud. I find this is a great way to collect feedback from developers with fresh perspectives who think up new, interesting and even crazy use cases for the platform. Ultimately these learnings are very important because they will help inform our strategy for future of Watson.
As a part of seeing such a high volume of hacks there will inevitably be a lot of uninspired, derivative, nonfunctioning, unrealistic and/or recycled junk out there, but I’ve hand selected five exceptional Watson hacks that I hope will inspire you. As with every hack, much of this stuff is buggy, to say the least — but that just makes it more fun after all, doesn’t it?
So in no particular order, and without further preamble, here goes:
The Story: Submitted for a Bluemix online hackathon, this project seems to solve the submitting team’s issue with finding answers they want — which are already readily available on irs.gov, but difficult to find and practically unsearchable. NOw you can use slack to get the answers you want quickly and easily — and IRS approved ;)
What it does: TaxSaver.club is a slackbot that uses Watson’s Natural Language Processing and deep machine learning to crawl irs.gov and answer practical day-to-day financial questions.
Watson APIs Used: Watson’s Natural Language Classifier
Other Services used: Slack, open source deep learning algorithms
The Story: Giflî was the First Place Winner of the Watson Developer Challenge online contest held via Devpost, an online developer contest website, out of the creator’s frustration with having to prepare pithy slide material for presentations.
What it does: Have you ever dreaded putting together powerpoint decks for every single pitch or meeting or presentation? Well now you don’t have to! Just talk, and Giflî will populate the screen with relevant gifs from Giphy’s database.
Other Services used: Giphy’s API
The Story: This is one of the first hacks I ran into that really blew me away. A winner at TheNextWeb’s New York City Hack Battle in 2015, Teachkit’s team was one person, used one Watson API alone, and was built in about 12 hours — all normally things that would limit a hack’s ability to be great, but this was definitely an exception. This is still the most interesting hack I’ve seen using Relationship Extraction. It’s an example of something that is not flashy, was not presented in a polished way and ultimately was not pursued by the hacker, but still has a lot of merit.
What it does: Teachkit is a mobile app that automatically analyzes written sentences and then visualizes and solves written math/logic problems to help students who find such exercises normally boring.
Watson APIs Used: Relationship Extraction
Other Services used: SpriteKit for iOS
The Story: This is a totally random hack using Watson and found via Hacker News. It’s also just ridiculously cool — built originally to create supercuts from The Wire, which while one of the greatest shows of all time in this humble writer’s opinion, is a very random and unnecessary task — but I digress.
What it does: Basically this is a service anyone can take and fork on Github to build an app which will automatically create supercuts from videos on YouTube based on a specified criteria. For example the creator has taken political speeches and created supercuts for some of the following topics: negative words, fighting words, past presidents and more. Please note: This is not an app, but rather a service that can be taken and configured to create supercuts around whatever a user desires. The applications are potentially very interesting and, it should be noted, hilarious.
Watson APIs Used: Speech to Text
Other Services used: Python, ffmpeg, and curl
The Story: Ok I’m cheating. Here are three kits the IBM Watson team put together to help folks get started building applications with interesting combinations of Watson services.
What it does: Let’s click the link and find out!
Watson APIs Used: Various
Other Services used: Varies
So hopefully this inspired some of you and helped spark some ideas. We’re always looking for more interesting projects so please if you have something or are thinking about something reach out to me or someone on our team . I’m on twitter at @Michael_Ludden and you can LinkedIn me as well.
- So yes it’s free and this is how: free use of the Watson Developer Cloud requires a free Bluemix account, which comes with a 30 day trial where Watson Developer Cloud services are free and have generally very high daily limits (no charge if you exceed the limit — it just stops returning your API call) and after the 30 day trial the Watson Developer Cloud services are still free — if you stay within the lower ‘free tier ‘ limitations for API transactions, which you can view through the Bluemix dashboard after you sign up (again, for free). I know… somewhat confusing but I gotta make sure I give this caveat. So yes it’s free and this is how.
The Story: EDIA was a runner up of the Watson Developer Challenge online contest held via Devpost, an online developer contest website, created in 24 hours to address the team’s desire to be able to quickly and objectively vet the realistic potential of a startup idea
What it does: You’re serial entrepreneur type that has new great startup ideas approximately once every hour, but how do you vet them all? EDIA automagically vets your startup idea using a ridiculous amount of Watson services to let you know if it passes what they call the ‘emotion’ test.
Other Services used: Ionic, Angular and SassBack for the dashboard UI, and Restify Node.js module to integrate various Watson services.