A blog series distilling QUANTITATIVE data concepts AND USE-CASES IN THE EINSTEIN ANALYTICS PLATFORM.
So let's say you have 2 fields in a custom object- ID and Estimate and you want to see how the estimate value changes over time. One approach would be to take a 'snapshot' of this data periodically --- daily, weekly,etc. Create a dataflow on a schedule and attach the following nodes to it.
1) create a data variable called 'snapshotDt' thru the use of a computeExpression node using the following formula > toDate(date_to_string(now(),"yyyy-MM-dd"),"yyyy-MM-dd") with the date format yyyy-MM-dd in the other drop-down boxes.
2) AFTER this node, create another computeExpression node and create a textfield of the same snapShotDt in order to make databindings in the dashboard simpler since you are dealing with 1 text field instead of three fields- year,month day. The settings in the computeExpression > text / SAQL expression toString(toDate('snapShotDt_sec_epoch'),"yyyy-MM-dd")
3) Make sure the object is initialized because the table will be recursively appended. Any misaligned fields will cause a dataFlow exeception. How? introduce an artificial date text for the 1st row then fire off scheduled flows to append to it going forward.