Import data via csv files (version 1.5)

You can download demo data here.

In order to load data to ForesightAI, you can either use one of our connectors if your data reside on the connected platforms or use this general import to include your data into the ForesightAI platform. The data should contain your product and sales information for optimum at least last 2 years in order to perfectly compute seasonality or any other trends.

This article describes the data structure exactly how we need it in order for ForesightAI to work correctly. Once you have the structure ready, you will be ready to upload the data and create a connection. If you are ready to do so, please read this article on how it is done. But first:

Data spaces

There are 2 data spaces for the ForesightAI.

• Sales Data

• Promotion Data

Sales Data contains all your product, warehouse and transaction information. Based on past transactions we can forecast future sales and based on product-warehouse data (current state of amount on stock for example) we can tell you what to order and when.

In order for perfectly analyse sales and predict future we need information about your promotions. Past promotions serve as a benchmark to see if the high sales are due to higher demand or due to promotion, future promotion serves us to higher the expected forecast because of it. In the best scenario both past and planned promotions will help us compute better forecast and orders.

Entity types

There are several entity types we import in order to compute all necessary data. Each fits into its own .csv file.

• Warehouse - this entity holds information about warehouse or any other place where we store items (point of sales etc)

• Product - basic information about product - name, category

• Supplier - basic information about suppliers

• Category - categorisation of the products (like product lines, brands etc.) - ForesightAI currently supports only one level of categories (so no nested or parent-child categories are taken into account)

• Warehouse-product - In warehouse product connecting data are specific data for each product on given warehouse (so for example product A on Warehouse 1 has different inventory on hand or price than on Warehouse 2)

• Transactions - Each transaction is one sales item, so if one order has 2 products sold, it is 2 transactions in ForesightAI. Data can be grouped per item per day (so all sold Product A per day from Warehouse 1)

• Currency - Indicating primary and (if applicable) secondary currency of the system to be displayed.

Data

Currency (currency.csv)

Warehouse (warehouse.csv)

Product (product.csv)

Supplier (supplier.csv)

Category (category.csv)

Warehouse product (warehouse_product.csv)

Transaction (transaction.csv)

Transaction types

In ForesightAI we can import multiple types of transaction (sales, purchase, stock-taking, transfer of goods etc.). The must for the integration are sales and purchase transaction. For explanation of the rest please contact us.

For the initial import, please consider using sales and purchase transactions - each has its own transaction_type_id.

1 = Sales Transaction (if the amount is <0 than it is Cancellation of the sales)

2 = Purchase Transaction (if the amount is <0 than it is Cancellation) - you are purchasing the product from your supplier

If you want to include cancelations (return of the goods) use the same transaction_type_id but use negative value in the amount field. The logic is always - positive amount = succesful transaction, negative amount = return or cancelation. ForesightAI does not calculate the cancelation exactly in the month as it occurred but tries to find the transaction which is cancelled and negate that previous sales. So that if you on 2020-31-07 sell 10 items and 1.8 ForesightAI see -10 items, you will see 0 sell in month 07 and 08 as well.

Read more information about transactions (and more deeper transaction types).

Replenishment

If you have multiple warehouses and want the orders to be generated only to the central warehouse please fill in the data as follows:

1) Central warehouse should be first in the warehouse file and its parent_warehouse_id should be empty

2) All stores which are replenished through central warehouse should :

• in warehouse.csv have filled in parent_warehouse_id as central warehouse

• for each item in warehouse_product which is replenished through central warehouse the "supplier_id" field should be empty

• for each item in warehouse_product which is supplied directly - supplier_id should be filled in

3) please make sure that each good in replenished warehouse there is the same product on central warehouse.

Optional Delivery Data

In order to tell ForesightAI that you are already expecting some goods (goods in transit) you can fill data for future_delivery. Please put in future_delivery only the goods which might come in near future. If the order was partially fulfilled we expect that the fulfilled part is in transactions as transaction_type_id=2 and the unfulfilled amount (should it still arrive) should be inserted into future_delivery data. The second type is customer order you are expecting in the future to be delivered.

Future delivery or Inventory in transit or Customer Orders (future_delivery.csv)

If you wish to give us during the integration phase your minimal periods (how often you buy such goods) and leadtimes (how fast you can receive goods upon placing an order) you can give us data in this format.

Supplier configuration (supplier_configuration.csv)

Later it is recommended to update the values in the ForesightAI application.

Promotion Data (promotion.csv)

Forecast only Sales data

If you need to compute only forecast and do not need to compute orders or help with stock, the minimal amount of information we need to get from you is this information.

The format of the files should be the same, the information not in the table below should be empty column or 0 for numbering fields.

Data Import CSV

Data are sent as CSV files. Entity type (entity type is for example Warehouse or Product) consists of all records of that entity type (so all warehouses are lines in one file warehouse.csv, we do not support multiple csv per type - for example it is not possible to send half of the products in one csv file and half in other csv file).

Foreign key - FK fields(connections between entities - like for example category to product link) must be filled and must match foreign entity, otherwise the row will not be inserted. For example if we find transaction which does not have product attached we ignore it for both forecast and orders suggestions.

Filename for each entity type is a possible to set up in initial setting, but should follow these naming conventions:

• warehouse.csv

• product.csv

• supplier.csv

• category.csv

• warehouse_product.csv

• transaction.csv

• supplier_configuration.csv

• currency.csv

Please also check the most common mistakes with data preparation when integrating ForesightAI with CSV files.

You can download demo data here.

Version

This document describes the data structure in following version:

• Version of sales data - 1.5

• Version of promotion data - 1.5

• Zendesk publish.4.log-707523872.zip4 KB