/Turnover-analysis-and-prediction

Turnover analysis and prediction

Primary LanguageJupyter NotebookGNU General Public License v3.0GPL-3.0

Turnover-analysis-and-prediction

Tasks:

  1. to make analytics according to the order history,
  2. predict the turnover for a certain period of time,
  3. estimate forecast error.

Stages of problem solving:

  1. deletion of 2013, as it relates to the time of the start of work;
  2. filtering data from incorrect or atypical;
  3. construction of a pivot table (turnover, the number of orders, average receipts by months with the division into primary and repeat orders);
  4. construction of seasonality indices (primary plus secondary orders) for turnover;
  5. trade forecast for the second half of 2017 by months and comparison with the fact;
  6. calculation of forecast error.

CSV file hear.

Brief verbal navigation on PDF file:

  1. pivot table is presented on pages 12-13;
  2. charts of monthly turnover, the number of orders and the average check are presented on page 16;
  3. the stages of the forecast are presented on page 17;
  4. a graph comparing the values of seasonality indices, calculated by method A and B, is presented on page 25;
  5. a graph comparing forecast values and actual turnover for the second half of 2017 is presented on page 28;
  6. a graph comparing the error values for the forecast, the seasonality indices for which calculated by method A and B, are presented on page 30.