LegosConsidering I grew up to become an engineer, it probably comes as no surprise that when I was younger the biggest highlight of every birthday was a new LEGO set. Each set gave me new pieces for my collection, and with every new piece came the possibility to build something entirely different. By the time I moved on to more durable construction methods, I knew how to work with gears, pulleys, belt drives, drive trains, and of course, spring-loaded exploding walls.

Like building with LEGOs, putting together a web application is a matter of combining the right pieces in the right configuration. Once you have established the overall purpose of your application, try to break it down into functional units. If you can find an API or library for each function, then all you have to do is figure out what you need to connect the pieces. This allows you, the developer, to focus solely on building something no one has built before.

To illustrate this building block approach, let’s dive into two web applications built by developers in the AlchemyAPI community.

Cruise Line Review Dashboard

Alchemy CruiseInteractive Dashboard • Created by Fabio Annovazzi

Alchemy user Fabio Annovazzi has created an interactive dashboard that highlights the keywords, sentiment, and subject-verb-object phrases found in cruise line reviews. The application stack starts by pulling reviews from Cruise Critic. These reviews are analyzed with AlchemyAPI’s keyword and relation extraction calls, with the addition of the optional sentiment parameter. The results of this analysis are passed to pandas, which helps pinpoint the locations in the original text of the extracted words. Finally, all the results are displayed in a number of interactive dashboards which were built on the Javascript library D3. The end result is a dynamic tool for uncovering insights and patterns in review data. Feel free to check out Fabio’s code, and start a conversation if you’re intrigued!

Yelp Data Visualizer 

Yelp Data VisualizerMore Info • Created by Balachander Ramachandran

Another interesting application of the AlchemyAPI technology is Balachander Ramachandran‘s Yelp Data Visualizer. Built at HackIllinois, the data visualizer creates word clouds that display keywords and sentiment from Yelp reviews. This project pieces together a number of powerful tools to build a flexible framework that can fit a wide range of data.

The data visualizer uses Flask as its web framework, with Bootstrap to handle the user interface aspects (for more on getting started with Flask, check out my previous post). The data navigation pieces are handled with Elasticsearch, AlchemyAPI is used for keyword and sentiment extraction, and additional language modeling is handled by the NLTK Python library. Sound useful? Download the code and give it a try!



____________________________________

Watson in the Wild is a weekly blog series that highlights cool apps that have been developed with Watson and AlchemyAPI services. If you spot an interesting reference to Watson or AlchemyAPI, or apps that others have created, please comment below.  If you’ve created an app that you’d like other people in the Watson Community to see, click here to share it with us.

 ____________________________________

The author, Zach Walchuk, originally posted this blog on The Alchemist Blog community site.

2 comments on"Watson in the Wild: Cruisin’ with AlchemyAPI"

  1. Thanks for sharing this article. I’ve been using Alchemy a lot lately, especially for making demos and teaching a programming course. I’m very interested in Deep Learning and AI development, and really would love to get into the ground floor at a company like IBM, working in relation to some of these projects. Do you have any tips or pointers for how to get more involved with this community?

    • Adam, glad to see an Alchemy user express interest in Watson. I’d point you to our Watson Developer Cloud offering as a good place to start. These are, like AlchemyAPI’s existing offerings, our REST API based services geared towards developers. If you have something cool you’d like to show the community we’d love to share it here like Fabio’s and Bala’s examples above. Post examples or links to your work in the forum.

Join The Discussion

Your email address will not be published. Required fields are marked *