For writers, having an understanding of how a story is playing out can be crucial to creating meaningful content. Using the Twitter-API, content creaters can create a custom search that show which aspects of a story are the most pertinent. Tweets are searched by topic or search term and tweets with a minimum number of likes (favorite_count) are included in the search results. This allows content creaters to quickly understand what topics audience members are most interested in at the time.
Search criteria also includes time parameters so that content creaters can view real time content and can create effective.
Searchterm: "Tom Brady", Timeframe: 1st hour of Superbowl game
The WordCloud library allows users to create generated word clouds as seen above. This particular cloud was generated by searching for "Tom Brady" in tweets tweeted during the Superbowl game itself. Using the Twitter-API, content creaters can create a custom search that show which aspects of a story are the most pertinent. Tweets are searched by topic or search term and tweets with a minimum number of likes (favorite_count) are included in the search results. This allows content creaters to quickly understand what topics audience members are most interested in at the time. The WordCloud library also allows users to fit words into a graphical outline image (above).
The outline image is generated automatically from an uploaded image. To create an effective word cloud, it is best to use an image that has a strong border around it. If there is not a strong border between the image and the background, words can “leak out” into the rest of the page. This was observed several times with different graphic images. Included here as another example using Alice in Wonderland as a background image.
Even with tools like Stopwords, a good amount of old fashion editing was still required. Spanish words krept in to the list and those needed to be taken out or added to the stopwords list. I chose to remove curse words from the wordcloud, but other users may find those four letter words very relevant to the story.
Twitter API - with credentials
matplotlib==3.0.2
numpy==1.16.1
pandas==0.24.1
Pillow==5.4.1
searchtweets==1.7.4
stop-words==2018.7.23
tweet-parser==1.13.1
wordcloud==1.5.0
A graph of most common words from tweets
some of the most common words from superbows tweets - oddball terms removed
RT @SJPeace: I heard the #SuperBowl is coming up.
3 of them killed 30 people. 3 of them killed no one. Look at the differences in the…Sun Feb 03 02:59:09 +0000 2019favorite_count: 26301
RT @SJPeace: I heard the #SuperBowl is coming up.
3 of them killed 30 people. 3 of them killed no one. Look at the differences in the…Sun Feb 03 02:59:09 +0000 2019favorite_count: 26301
RT @myserenity69: It's on #SuperBowl Sunday once a year..
BUT RACISM IS A DAILY REALITY.
Remember to #TakeAKnee
BlackLivesMatter #No…Sun Feb 03 02:59:10 +0000 2019favorite_count: 491
RT @tribranchvo: straight men: are you watching the superbowl this weekend?
me: https://t.co/yvSPyJu87gSun Feb 03 02:59:10 +0000 2019favorite_count: 32199
RT @SJPeace: I heard the #SuperBowl is coming up.