The data then goes through a validation and cleansing process before being submitted to the Retailigence Machine Learning server. The steps are:
A. Data Format Validation
This step validates the format of the input data. If the input data is not in the format expected, this step gives an error and turns red. Clicking on the red icon will display the data fields which are incorrect.
Please contact a system administrator or raise a ticket to Retailigence to get the data issue resolved.
B. Data Cleansing and Data Preparation
This step indicates that the data is being cleansed. Any data elements which are considered ‘noise’ are cleaned out from the input data. There are two threshold points to be considered
- Number of units sold across the business – If this is below a certain number then those items are filtered out.
- Number of units sold in any store – This is a smaller number than above. Item -locations below this number are filtered out.
This process could fail entirely or pass with certain items filtered out. If any items are filtered out then the Stage in the process flow will turn amber and clicking on it will show the items or item / locations filtered out.
Note: The data cleansing is only meant to remove ‘noise’ from the clustering algorithms. The sales data will be added back for the assortment optimisation run.
C. Data Processing
This step indicates that the machine learning algorithms are being run.
Failure at this stage is due to a lack of data through a small product group or time selection, causing the algorithms to not have enough data points to produce a suitable output.
D. Store Clustering
This step indicates that the Clustering Dashboards are being generated.
Downloading Data
At any stage you see the symbol 
The user can click it to download the data for further analysis. This downloads a zipped excel file containing more detailed data regarding the selected screen.