Customer feedback on social media can help businesses make a difference in their sectors of activities. We thought of an easy way to make every single business enjoy the benefits that come along with social media presence. Our solution, B2C Analytica, is a data analytics platform that performs sentiment analysis on data collected from social media using artificial neural networks and other natural language processing techniques. We do this help businesses make the right decisions in their day-to-day management activities.
What it does
B2C Analytica uses user comments collected on social media to predict and monitor user behaviour on social media. It uses a recurrent neural network model trained with PyTorch for sentiment analysis on data collected on social media. The predicted sentiments are then used to monitor customers and determine business performance with the help of tables and graphs.
How I built it
To build this solution, we went through the following steps.
- Requirement gathering.
- Collecting data on social media to form a dataset for the model
- Training a sentiment classifier model using Pytorch
- Building a web application
- Connecting the trained model to the web application to form the final solution
Challenges I ran into
- This solution requires an inclusive dataset to work perfectly which wasn't easy to get to we used a sample dataset collected on the internet to train the sentiment classifier model.
Accomplishments that I'm proud of
No matter the challenges, we built a prototype for our solution.
What we learned
Getting the right dataset and training the model is not as easy as I thought.
What's next for B2C Analytica
We are planning to deploy our solution to help small businesses kick-start without necessarily running and ads ( advertisement ) campaign. Thereby saving money. Also, we would like to build a Facebook messenger bot in the future.