This project intends to provide a comparative model and data analysis report which could be helpful in weather-dependent resource planning. We have used two predictive algorithms namely Linear regression and Regression Tree. For trend analysis, we have implemented a two-fold analysis approach based on Months and Station locations using two non-parametric methods namely, Mann-Kendall method and Sen's Slope Method.
The DSWR Process Notebook.Rmd describes the course of the project. This also includes deviations from the initial proposal and further information about the evaluation of our project.
The website can be found at https://sites.google.com/view/dswrweather/home
The screencast can be found embedded on the website mentioned above or can be accessed directly on YouTube: https://www.youtube.com/watch?v=jZp-fKNxvwc
###Data set used:
Step 1) The main dataset used for our analysis Weather_dataset.xlsx.
Step 2) We have used FinalModifiedData.xslx file for our project
Step 3) For Station-based, predictions as well as for trend analysis we have used different files that all are present in the folder Station code
We have different files for predictive and trend analysis, below are the details:
- LinearRegression.Rmd : contains the code and logic behind the implementation of the model along with the results.
- RegressionTree.Rmd : includes all the code used for creating the different models and their evaluation results.
- TrendAnalysis.Rmd : includes all the code used for trend analysis based on stations and their results.
- TA_Month.Rmd : includes all the code used for trend analysis based on stations and their results.
- TA_Result_Viz.Rmd : includes all the visualizations used for trend analysis
Note: Download all the rmd files and their respective input read files and put them in the same path location.
