Fake News Analysis

Inspiration

We were inspired by the opportunity to employ our knowledge to tackle this issue in the most effective way possible through programming and teamwork. As we aspire to be data scientist in the future, we though that this could be a really good way to learn about this field and put our abilities into practice.

What it does

We coded a lot of functionalities. First we start by performing a simple data processing so it's easier for us to manage our data. Then we learned about what does it contains. Then we started by doing a sentiment analysis to search for differences between fake and real news and saved in a new column. Also generated two new features called text len and word count. From then we plotted this so we can look up for interesting insights or revealing information. By the end we do a word count so to search for common words in both types of news, taking into account that some words are useless like stop words, and also using tokenization or tagging.

How we built it

We were able to perform this analysis using the Python programming language and libraries such as NumPy, Pandas, TextBlob, among other tools that helped us to analyze the data and compare results in the end. Many of them were very new for us so we learned how to use them.

Challenges we ran into

We encountered several challenges, such as finding patterns within the collected information and relating it to our purpose of determining whether a news article tended to be false or true. We also faced coding challenges that we overcame through extensive research and mutual effort between my partner and me.

Accomplishments that we're proud of

We are proud to have learned more about these types of challenges and to have enriched our knowledge. It also helped us a lot to learn about new tools and methods that we can implement in future projects.

What we learned

We learned to generate a comparative analysis between two sources of information using different tools provided by the python language, and in this way, to be able to search for solutions that contribute to this objective. A lot of new things from natural language processing were very new, we never heard of them before. So having to search for it was very great and now we can understand how great analysis and artificial intelligence models work.