Inspiration
Nowadays, there are lots of news everyday. Many of the news are very long that take a lot of time for people to read about. And the need for quickly knowing the hot topic is growing. Thus, we want to develop an app to automatically summarize the news so that users can look at news for shorter time. We exact the keywords to build tag cloud so that people can know what are the hot topics everyday.
How it works
We build the end-system to get the feed news from BBC, CNN, New York Times and other websites. We use python to exact the main text in the HTML and summarize it by AI algorithm. Meanwhile, we can exact the keywords and calculate the "hot degree" in the passage. We use mongo-db to store the data and build a ruby-on-rails server to send the data to IOS. In the front system, we build the app based on IOS. We use some really cool interface to make sure that user can have better experience.
Challenges I ran into
In the end-system, it is hard to exact keyword and summarize the text due to lacks of time and knowledge in the AI field. However, we handle the challenge by quickly implementing AI algorithm based on existing papers and libraries. The front-system is hard, too because we want to implement waterfall, cloud tag and other effects. However, we handle all the difficulties by ourselves.
Accomplishments that I'm proud of
I am so proud of our team work. We four are classified as 2 sub-groups: front-system group and end-system group. We collect the data and search for the API through internet. We refer for some existing projects and implementations in some of our algorithms. I am proud that we can handle all the difficulties by ourselves.
What I learned
We learn a lot about how to write a ruby-on-rails server, exact keyword and produce summarization automatically based on AI algorithm. We learn how to implement an IOS app with some really cool effects like waterfall and cloud tag. We learn how to manage data, how to scale out data and to send request and response between front-system and end-system.
What's next for Easy News
Due to time limitation, we just work out a demo version. We should do more work on the following things:
- scale out the data in mongo-db.
- we should improve out AI algorithm to exact main text, keywords and producing summaries.
- we should deploy our end-system online so that the service can be used for day long.
- we should do some improvements in IOS effects and effeteness.
Log in or sign up for Devpost to join the conversation.