Sales forecasting Model

• 

  A plot of the raw data. Black dots are the actual closing value and blue line is predicted closing value

• 

  Main page where the predicted raw data is displayed

• 

  More components to help users understand the trend better

About

In my opinion ML/AI is the future especially for organisations, hence I have made an user friendly web app which uses ARIMA model powered by FBProphet to predict the price/closing value of an asset/good/service. The main purpose of this model is to help businesses/organisations to make the most out of ML models easily! The ultimate goal is to make a platform in which the user can have various option of selecting ML/AI models based on their accuracy for prediction.

What it does

The model can predict the closing value of any good/services/assets based on its historical data (.csv). It only requires two columns of Date and closing value in order to forecast. This model essentially helps organisations to set their plans/goals and help with anomaly detection.

Tech Stack Used

• Python
• Streamlit
• FBProphet
• Pandas

Challenges I ran into

The most challenging part was to integrate ARIMA model with the web app as both had different set of rules

Accomplishments that I'm proud of

Making the model and the web app public

What I learned

This project surely exposed me to new opportunities like expanding the web app and making it public!

Future Plans

• Add 'file upload' feature to the app so the users can directly upload their historical data
• Add more editable parameters for the users
• Integrate more models like LSTM

Built With

• fbprophet
• pandas
• python
• streamlit