HOW WE DID IT
To achieve the results State Farm required, our team worked on
different aspects of the analysis and research, as well as the
creation of the website. Early in the semester, Alex Carney got an
early start on writing a program using the coding language Python,
that retrieved, processed, cleaned and analyzed tweets that had to do
with State Farm. It was important that we got this code working
quickly and correctly because this code allowed our team to find and
store all of the tweets that we would need for analysis. Alex used
various Python libraries including Textblob, GetOldTweets3, and
Preprocessor. Alex had to piece together these libraries in order to
create his own script that was not bound by the limitations of the
Twitter API, which limits the amount of tweets that you can pull off
of Twitter in a certain amount of time. These libraries also allowed
our team to account for emojis in tweets, which is not something that
is usually considered in sentiment analysis. We felt that adding
sentiment scores to emojis not only set us apart from other sentiment
analysis, but enabled us to have a more accurate view of the overall
tweet. In total, our program retrieved upwards of 133,000 tweets that
were about State Farm. Once we had accurate and clean data to work
with, Amanda Delargy worked with Python and R program to create an
interactive graph with a range slider that shows the public opinion
of State Farm over a set period of time. We felt that having an
interactive graph would be beneficial to the person interpreting it
because it offers a deeper understanding of the material as well as
making the webpage more interesting instead of just text and
pictures of graphs. Amanda used the Python libraries known as Pandas
, Matplotlib and Plotly. This allowed her to turn our CSV file full
of tweets into a dataframe so that it could be more easily
manipulated. This dataframe was then turned into a line graph and
made interactive by utilizing the Python libraries previously
mentioned. Emma Bradley, Yonny Sohn and Jake Walter then took this
information and created a website, with Emma and Yonny focusing on
the back end development of the website and Jake focusing on front
end. By delegating different responsibilities to different members
of the group, our team was able to finish our project in a timely
manner. Our team did not have many problems during this project
because of clear communication as well as reaching all of the goals
that our group had previously set for themselves so that we would be
on track and not fall behind. Overall, our team created a useful and
interesting way to showcase customer feedback for State Farm.