Time-Series-Analysis

This is basic evaluation exercise on open sales dataset. Please follow the steps. Fork the Notebook and commit the solution. Share your git with solution with TA

Note: please use markdown for highlighting the below solutions.

Business Questions & Descriptive Statistics

Which Category is Best Selling and Most Profitable?
What are the Best Selling and Most Profitable Sub-Category?
Which is the Top Selling Sub-Category?
Which Customer Segment is Most Profitable?
Which is the Preferred Ship Mode?
Which Region is the Most Profitable?
Which City has the Highest Number of Sales?

Fork the git

Preparing the Environment

Jyupter Notebook
Pandas — Data manipulation
Matplotlib and Seaborn — Data visualisation
Time series analyasis & Decomposition - use statsmodels library

Data Cleaning

Handling Missing Values
Duplicate Data

Correlation betweeen 2 features

High +ve correlation
High -ve correlation

Univariate Time Series Analysis - Forecasting future sales (Furniture or Office Supplies) according to shipping date using ARIMA

Validating Forecasts

Train the model until 2016 and use year 2017 for forcasting

basudev2016/Time-Series-Analysis