Tagging System

The Tagging system is a web application that automatically assigns tags to stack overflow questions based on their content.

Benefits of an autonomous tagging system

- Improved discoverability of questions : When questions are tagged correctly, they are more likely to be found by users who are lookin for help with similar problems
- Increased relevance of answers : When answers are tagged with the same tags as the questions they are answering, they are more likely to be relevant and helpful
- Reduced workload for moderators : Moderators can focus on other tasks, such as reviewing questions and answers, when an autonomous tagging system is in place

Team Members and Roles

Team Name : Tag Masters

1. Roopa Dharshini - Teal Lead, ML Developer
2. Harish Raghavendar - Cloud Developer
3. Ujjwal Anand - Back-end Developer
4. Hritesh Sinha - Front-end Developer

Uniqueness of Tagging System

• Introduced Copy-to-Clipboard button
The copy to clipboard button is a small button that allows users to copy tags to the clipboard. This can be useful to copy tags from 'Tagging System' and include those tags while asking questions in Stack overflow website, so that the user can found solution easily

How to access the web app ?

To access the Tagging System web app, follow the instructions below:

1. Ensure you have Python 3.x installed on your system

python --version

2. Clone this repository

git clone https://github.com/Rupa-Rd/SBSPS-Challenge-9913-Autonomous-Tagging-Of-Stack-Overflow-Questions.git

3. Navigate to the project directory

cd auto_tagging

4. Install the necessary modules

pip install -r requirements.txt

5. Run the app from the command line

python manage.py runserver

6. Launch the app
Find the app in port 8000

App Screenshots

• Landing Page / Home Page



Example for Python Tags

• Question & Title Input Page


• Tags Prediction Page


• Tags Copied to Clipboard



Example for .Net Tags

• Question & Title Input Page


• Tags Prediction Page



How it Works ?

1. The user should provide the Title and body of the question in the landing page of the "Tagging System App"
2. The title and the body of the question will undergo vector transformation before feeding the data to the trained model
3. The App access already trained model in `model.py` as a pickle file inorder to make predictions
4. The predicted tags will be a set of numeric arrays, this then undergo inverse transformations inorder to convert it as a `string`
5. Then the predcited tags will be displayed in the web page which allows the users to copy and use it

Tech Stack Used

• Model Preprocessing
• IBM Watson (NLP)
• Pandas, Numpy
• Scikit-learn
• Model Training
• Support Vector Machine Model
• Logistic Regression Model
• OneVsRest Classifier Model
• TfIdVectorization
• Front-End
• Html
• Css
• Javascript
• Back-End
• Django
• Python

IBM Cloud Technology

• IBM Watson Studio
We have used IBM Watson to pre-process the Stack overflow dataset given to us. We used 'NLP module' in IBM Watson studio to remove the following :
• To remove punctuations (,.?/:""''*&)
• To remove stop words (a,an,not,..)
• To replace words (don't - donot, can't - cannot, ...)
• To drop the unwanted columns (Score, Date, Time)
• To check and replace the null values

Advancement

1. Preparing to deploy the app in the cloud
Our app will be available to everyone, regardless of their location, by deploying it in the cloud
2. Improving the predictability of the model
The current model has a 60% prediction rate; with more data, it will be able to predict a larger number of tags with better accuracy

Confused?

If you're confused in accessing the "Tagging System" web-app . Watch this video. https://drive.google.com/file/d/13quGbU26y3NOZu0nghHuTwW80IEfZGPJ/view?usp=sharing
• Click here to watch



The app is all yours. Have fun!😊