Transactions
Where your data varies over time, and you track it on a granular basis
The Transactions standard data format allows you to pull in data on a transaction-by-transaction basis into Causal. This is useful for things that are recorded on a per-day, per-week, or per-event basis, like Sales by Customer.
Causal allows two types of Transactions format.
Standard Format
The Standard format rules are:
- The first column must have the name "Date"
- Subsequent columns can be Data Items or Dimensions with whatever name you choose.
- You can follow a column name with [dimension] or [variable] to ensure Causal recognises the column correctly
Say you're pulling in the above example data into a monthly model, Causal will add all of the transactions for each month, according to the date on which they occurred.
You can add additional columns of Data Items or Dimensions to the right. See example screenshot below. You can also add mappings between Dimensions. The column name must have the name of the dimension being mapped from and the name of the dimension being mapped to, in the following format: [from dimension] > [to dimension]
So in this case we are bringing in an additional Dimension Account Manager which is mapped 1:1 from Customer.
You can also add a Cohort column if you want to bring in cohorts!
Alternative Format
Alternatively if your data source has Data Items coming through in rows rather than columns then you can use the Alternative format for data ingestion.
The first column must have the name “Date” The second column must have the name “Type” and contain the Data Item/ Variable names. The third column must have the name “Value”.
Please note that if you want to add a Cohort dimension this column must be directly after the Value column.