Back to Parent

Outcome


Intention

When first trying to come up with ideas for a chatbot, I initially thought of the tomigatchi. As a kid you would take care of it and it would be your companion that you would basically have as a virtual pet. During this project that was my inspiration behind SubWoofer the dogbot. I wanted to create a simple yet fun bot to interact with that would simulate talking and playing with a dog. I thought this would be a great and fun way to use different APIs to have the dog be a friend and helper.

A target audience I was thinking about would be kids, because many want pets, but many do not have them. And let's be honest kids these days love playing with their parents phones whether its apps or something else. This would be a great way to distract the kids while giving them their longed for pet.

Context

Like above I drew inspiration from the tomigatchi phenomenon back in the day. I also reviewed the twitter account thoughts of a dog to think of how it would be fun to speak like a dog, spelling errors and such. I imagined the personality of my bot to be that of doug from the movie UP. when doug gets a talking collar on it, he is a fun loving, dog who is just excited to be with humans. He has his silly tendencies, including stopping conversations when he hears the word 'squirrel', and ultimately just wants to be loved.  

Process

My main approach was to start with the conversation aspects first. Since this is a dog and there isn't just one function of the bot, I had to take into account that there could be many dialogue options. This was a very interesting and difficult task to take on. I needed to throw in many keywords and after every message output I kept coming up with more and more ways the dialogue could go. 

After I came to an acceptable, yet still in need of work, amount of responses I began to work on the giphy calls. Since this bot is playful it would only be fitting is it sent gifs with the responses. I used the giphy API to determine gifs and even search gifs when the tricks option was given. Many responses had an emotional aspect to them and using gifs it is easier to view what the dog was feeling. 

Another API I used was the twitter API. This allowed me to call upon a certain users tweets and push them to the text message. One thing I could not figure out was how to only get the tweets and not the replies from the timeline. Also using this API, the dogbot was able to send a tweet for you by performing a command. 


Keywords
Show Advanced Options
Img 1294
Show Advanced Options
Img 1295
Show Advanced Options
Img 1296
Show Advanced Options
Img 1297
Show Advanced Options
Img 1298
Show Advanced Options

Reflection

Over the course of this project I was able to apply a lot of what I learned. From implementing responses and using APIs to just coding a bot in Ruby. The twitter API was new because I had to basically apply for it and going through a website as big as twitter was pretty interesting. 

Due to time constraints this bot could not reach its full potential. In the future this could be even more conversational utilizing even dialogue flow and other resources to make the conversation seamless.  Also in the future this bot could use more of the functions of the twitter API, it can send tweets now but could also help follow, favorite and retweet other tweets the user finds interesting. One item that I tried using was a dog breed API to give information on different breeds and their backgrounds. This could be used to help inform the user about different breeds and if kids are using it then could help identify which breed they can match with.

Show Advanced Options
Drop files here or click to select

You can upload files of up to 20MB using this form.