In this project, I have drawn various inferences from the yearly sales of eletronic products.
The dataset was a combination of 12 csv files representing the yearly sales of the products. Some data cleaning was required as well as the datetime column didn't had the correct datatype. So, all these were performed while combining the individual dataframes itself. The concatenation keys were inferred from the filenames using regex ;).
Then various inferences were drawn from the dataset:
- Finding the best month for sales (Fig 1)
- Which times are best for advertisements (Fig 3)
- Which combination of products sells most
- And others...
With growing data the need for data wrangling becomes evident. This project is an analysis on the overall sales of electronic products over the whole United States. The main objective was to find the business-interested topics. The general user was observed and it was seen that the interests and are similar across times as well as cities with some peak times of sales.
(Originally posted on youtube)