Mad Lib-erator

the classic. the original. mad libs!
when inputted into a mad lib, the keyword becomes strangely relevant
so relevant that the different parts of speech create the theme of that story!

Inspiration

Do you remember all your childhood memories, now tucked away in dusty boxes in your attic? And, perhaps, inside one of those boxes are stacks of Mad Libs, completed with friends and family during long car rides and rainy afternoons. Mad Libs, a classic American word game invented in the 50s, consists of one player prompting the other player(s) for random words that complete the sentences, resulting in a nonsensical story. Throughout the years, there have been at least thirty different themed versions, reinstating their mark on American culture with each release. However, what if the user could determine the theme of the story rather than the story determining the topic? Our team explored this possibility of rebooting the American classic with a new, modern approach.

What it does

The Mad-Liberator helps despairing Mad-Libbers by prompting the user to input parts of speech and a keyword to serve as a 'theme'. The Bing web search API uses the ‘theme’ to extract text from the first ten search results on Bing. Using this as source text, and the Linguistics API for part-of-speech recognition, the Mad-Liberator randomly selects words that match the part(s) of speech the user asked for.

How we built it

The Mad-Liberator was programmed in Python and implemented the Microsoft Bing web search and Microsoft Linguistics APIs.

Challenges we ran into

We started out with a different idea -- we were going to learn about neural networks and write code to search the web for images, merging them to make a new, generalized image for a keyword

Fun fact: We don’t know anything about neural networks
We don’t know how to line up images to merge semi-transparent images effectively
Changing ideas with less than 24 hours left until project submission! B-)
We’ve never used these APIs before
How does this work? --> I just want to search Bing for a dog
I miss my childhood and games
I miss playing Mad Libs!
Newer APIs = less documentation
#saved by asking for advice from fabulous Microsoft hackathon volunteers (thanks Microsoft-Clemens!)
Merged our code for the Bing Web Search API with our code for the Linguistics API
Wrote a coherent project description in time for submission

Accomplishments that we're proud of

being able to integrate different APIs into one coherent project
connecting our reminiscent pasts with our technology today
learning how to use APIs, since both of us had little experience beforehand
learning a lot! (from our mistakes)

What we learned

We learned a lot about what Microsoft Cognitive Services APIs do and how to use them and integrate them into a project (especially the Bing Web Search and Linguistics APIs). We also learned a great deal about neural networks for the first half of the hackathon (oops).