The solution for evaluating data in real time which help business perform better and keep track of all relevant content about a brand in real-time, perform analysis as topics or issues emerge, and detect anomaly with alert. By monitoring brand mentions on Social media brands could inform engagement and deliver better experiences for their customers across the world.

What it does

An end-to-end real-time Twitter monitoring system to track customer behaviors towards brands by identifying sentiments fluctuation, analyzing trending topics and geographic segmentation, and detecting anomaly on scandals, with consideration of improving brands customer engagement and retention.

How I built it

  1. Extract streaming Twitter Data, preprocess data in Python, and load data into MySQL for storage
  2. Perform exploratory data analysis using Pandas, NLTK, Seaborn to explore the insights
  3. Connect with Plotly for real-time interactive dashboard based on time series

Challenges I ran into

  1. Unstructured tweet texts may contain messy code and emoji characters
  2. Some brands may take a long time to collection enough data to perform analysis on issue emerging since they target specify groups of people
  3. Some hot words will uncover useful insights only after appearing more than 10k times on tweets
  4. Plotly doesn't have well-document on reference making customize dashboard much harder

Accomplishments that I'm proud of

I'm proud that I was able to do data analysis in real time I'm proud that I was able to put forward my ideas and come up with on spot decisons

What I learned

I learned how Natural Language Processing works. Learned how to allocate work to team members and make sure everyone is on the same page.

What's next for Real-Time-Data-Analysis-of-Social-Media-Posts

  1. Let customers know how people are feeling about jetblue in real time while booking tickets.

  2. Immediately send alert messages to the jetblue staff in real time and respond accordingly.

  3. Search for tweets in all languages.

Built With

Share this project: