Data Import from External Sources into FoodChain-Lab

Content

  • This tutorial shows how data can imported into FoodChain-Lab without using the Excel templates.
  • Therefore the external data has to read into KNIME tables.
  • KNIME provides a multitude of reader nodes for various data formats (e.g. csv, SQL databases, …).
  • Here we just show which tables are needed by FoodChain-Lab and which columns they must contain. To do so the we have created these tables with Table Creator nodes.
1

2
  • FoodChain-Lab needs three tables to perform a tracing analysis: Stations, Deliveries and Delivery Relations.
  • The Table Creator nodes on the left show all mandatory columns for these three tables.

3
  • Double click on the Table Creator for the Stations table to open its dialog.

4
  • As you can see the only mandatory column in the Stations table is the column ID of type string.

5
  • Double click on the Table Creator for the Deliveries table to open its dialog.

6
  • The Deliveries table has three mandatory columns: ID, from and to (all of type string)
  • The from and to columns contain the source station and target station IDs.

7
  • Double click on the Table Creator for the Delivery Relations table to open its dialog.

8
  • The Delivery Relations table has the columns from and to of type string.
  • These columns contain IDs from the Delivery table and are meant to connect two deliveries.
  • The from-delivery is a part/ingredient of the to-delivery. A contamination can spread from “from” to “to“.

9
  • Now we will look the Table Creator nodes on the right which show the optional columns for the three tables.

10
  • Double click on the Table Creator for the Stations table to open its dialog.

11
  • The optional columns for the Stations table are Address, Country, Latitude and Longitude.
  • Latitude and Longitude are used for the geographical visualization in the Tracing View and Address and Country can be used for geocoding, if Latitude and Longitude are not known yet.
  • Any other arbitrary columns of this table can also be used in the Tracing View for highlighting etc.

12
  • The Geocoding node in the upper right corner shows how Address and Country can be used for geocoding to compute Latitude and Longitude.

13
  • Double click on the Table Creator for the Deliveries table to open its dialog.

14
  • The optional columns are Date Delivery, Date Delivery Arrival, Lot ID, Amount and Amount Unit.
  • The date columns are of type string and in YYYY-MM-DD format. They used for computing cross contamination.
  • If the Lot ID column (of type string) is provided, lot-based scores are computed.
  • The amount columns are just used for plausibility checks.
  • Any other arbitrary columns of this table can also be used in the Tracing View for highlighting etc.

15
  • Double click on the Table Creator for the Deliveries Relations table to open its dialog.

16
  • The Delivery Relations table does not have any optional columns.
  • All columns other than from and to are ignored.