This project aims to predict the temperature at New Delhi, India using various time series concepts.
The dataset contains datapoints from the year 1996 to 2016.
An in-depth analysis of the project is given in Time_series_report_Udbhav_Kush.pdf file with all the results of the code.
The main project file is the Delhi_Weather.py file. The toolbox.py contains all the supporting functions for the project which have been developed throughout the course duration.
We just need to run the Delhi_Weather.py to run the full code.
A small dashboard has been created for the dataset to analyze the different parameters in the dataset visually. The dashboard code file is app.py.
P.S. The LSTM code sometimes gives Nan in the training loss and validation loss, which will lead to error in the file. In that case, we just need to rerun the file and the file will be running.
Please refer to the medium article of this project to know more about it: https://medium.com/@udbhavkush4/delhi-weather-prediction-using-time-series-1ebe509cf73