TwitterAggregateSolution

Inspiration

I decided to make this project because I thought, what if I could acquire a rough estimate of what people are feeling about this keyword right now, within a minute? And I wanted to work on a major API for this Hackathon, henceforth I chose to create a project that uses aggregated twitter feed.

What it does

First, using my dev API Keys, I can return a search query on a specific keyword through the Tweepy api, which is a JSON. Then, using Pandas Dataframe I cleaned the data and returned the only the text part of the tweet. Before this, I have already found a dataset for training the logical regression algorithm for text, so I used that to predict and return a meaningful data.

How I built it

I built it using Jupyter Notebook, the hallmark of data science platform, and used GitHub to store the project.

Challenges I ran into

Quite a few actually.

First, I worked on this project alone. Now, I'm not a lone wolf type person or anything, but it was something I wanted to try for this Hackathon. This did not turn out as effective as I thought it would be first. I did not have enough expertise to take care of the frontend and the backend of the program under my allotted time, and when I got stuck on a problem, I did not have someone to brainstorm with to come up with creative ways to solve that problem. That's one of the reasons why there's only a Jupyter Notebook is in the GitHub repo right now.

On the more technical side, I had problems with using regular expressions, and cleaning a messy data proved harder than it was. There were many other issues with data types that prevented me from doing what I wanted to do, which slowed my progress down significantly.

Accomplishments that I'm proud of

Everything said and done, I did reach my end goal in terms of project. Regardless of how frontend looks like right now, at least that's something I can implement later on. Also I attended every workshop possible to gain insight on my project, and that's also something I'm proud of.

What I learned

It is very inefficient to work alone. I need to have a teammate that complements my skills whom I can clearly communicate with. I need to improve my personal skills as well. I significantly lack frontend skills. I will dedicate myself to learn HTML/CSS, Javascript, and React. I need to improve my knowledge on how to use data science related functions: regex, dataframe, logistic regression, etc Pico-8 is really awesome. With this in mind I'm going for the win next time.

What's next for TwitterAggregateSolution

I'm going to conduct more tests to see if the trained model is actually working for the aggregated twitter data. Then, I'm going to create a frontend for the project in a web environment.

Built With

• jupyter-notebook