Inspiration
In today's digital world everyone is very active on social media. Starting from personal posts up to companies commercials or news sharing, hashtags have become the way to efficiently spread information social networks. However, some hashtags are more popular or suitable for specific topics than others and finding perfect description is not always easy. Our idea was to make a Messenger bot that would enable easy posting of photos together with relevant hashtags and description on Twitter.
What it does
Messenger bot enables the users to upload a photo and then get a number of suggested hashtags based on image content. The bot is able to suggest tags relevant to objects, people and emotions that are present in the image. Furthermore, it will detect if there are famous landmarks or celebrities in the photo that the user would like to post about.The bot is also able to provide a couple of funny picture descriptions based on the content of the image.
How I built it
Messenger bot is built using the Microsoft Bot Framework. The uploaded image is sent to Microsoft Azure Cognitive Services for processing. We are using Face API and Vision API to extract all user relevant information from the picture that would be interesting as hashtags and image description.
Challenges I ran into
Waterfalls in NodeJS were not that fun to play with. Microsoft Azure Cognitive Services picture description describes the picture from third person, not from first person, so we had to personalize and adjust our messages from the provided tags.
Accomplishments that I'm proud of
Integration of many different services in our app was real challenge that we are happy to have accomplished in an efficient way and made something fun, easy and useful in a short amount of time.
What I learned
We have learned how to easily integrate the powerful service provided in Microsoft Azure in order to make interesting and helpful tool for social media making life easy for all those tweeting fans out there (and we all know they are many).
What's next for Twizzy
Features for hashtags that Twizzy now provides are just a start. We would have like to also add options like finding funny sentence combinations from selected hashtags and integrating them into our servicce. Our bot at the moment also randomly selects number of tags to suggest from multiple provided hashtags, but the nice feature would be to check how popular each tag is and then suggest the most popular ones. Also, it would be nice to make it possible for this bot to post on different social networks like Instagram and Facebook in addition to Twitter. The popularity of hashtags should be optimized for each network specifically.
Log in or sign up for Devpost to join the conversation.