This is a B2B sales prediction problem. The dataset available (includes a mix of synthetic and real data) is from a steel manufacturer that has other businesses in the Auto, Metal Fabrication, and Infrastructure as customers. The goal is to predict the quarterly sales to each of the 75 customers. In addition to the company-specific data, general economic indicators are also included that can be used for the purpose of prediction.
The goal is to predict the quarterly sales of various customers of a steel manufacturing company.
The final results are measured using Mean Absolute Error (MAE) evaluation metric.
I have used Jupyter Notebook, you can use google colab as well.
Data Preprocessing: I have used comprehensive data preprocessing techniques like Data Cleansing, Handling Missing Values, Handling Outliers, and Handling Categorical Data.
Data Visualization: I have used box plot to visualize the outliers.
Regression Model: The model I have used here is the XGBoost Regression algorithm to get the accurate quarterly sales of the company.
Evaluation Metrics: The model's performance is measured using Mean Absolute Error.
The dataset contains the following files:
EconomicIndicators.csv: Supplemental economic information for the corresponding period in train/test data.
train.csv: The training dataset.
test.csv: The test dataset.
sample_submission.csv: A sample submission file in the correct format.
train/test.csv:
ID - Row id column
Company - Name of the company/customer
Quarter - Quarter for which the sales are provided/to be predicted
QuickRatio - Financial ratio indicating the customer's liquidity situation
InventoryRatio - Ratio of sales over inventory
RevenueGrowth - Revenue growth projections based on analyst and company projections
MarketshareChange - Market share growth projections based on analyst and company projections
Bond rating - Bond rating of company
Stock rating - Stock rating of company
Region - Region in which the company is situated or operates primarily
Industry - Industry are of company
Sales - Sales for the given quarter (target variable)
EconomicIndicators.csv:
Month - Month for which the indicators are provided
Consumer Sentiment - Consumer sentiment index value based on survey of consumers
Interest Rate - Average yield of 5 year US Treasury
PMI - Purchasing Managers Index
Money Supply - M2 Money supply
NationalEAI - National Economic Activity Index
EastEAI, WestEAI, SouthEAI, NorthEAI - Regional Economic Activity Index
You can reach out to me on LinkedIn for any questions or feedback. Click on the LinkedIn Icon to land on my LinkedIn Profile.
