Creating Timeline from Start and End Dates Columns

[intapiuser]

Introduction

This article will demonstrate how to quantify on a date axis the availability \ unavailability of products given only the start and end date of each product's availability (See data format in "ProductData,csv" table below).

Business Case

Viewing how many campaigns were available on a collection of dates, how many machines were running or how many products were available in a collection of dates. 

Example

Steps

To Achieve this we will be creating 3 tables in our ElastiCube:

1. ProductData: (ProductId, StartDate, EndDate) Source table, containing for each product its ID and availability date
2. AllDates: (ProductId, dates) Custom table, holds a Cartesian product of all dates given (StartDate and EndDate values) multiplied be all products
   
3. Availability: (dates, avail, notavail) For each date it holds the number of available products and the number of unavailable products 

1. Add your data in a similar format of ProductData. Make note of the column names as we will be referring to them in the next steps.
2. To create AllDates table either:
    - Import a table into the elasticube containing all dates in the rance and rename it to "AllDates" - this is the preffered method
   OR
   - Add a custom table with the following dynamic script (changing table and column names accordingly) - this is the less-preferred method as it may result in 'holes' in the timeline if some dates values are not imported originally
   
   SELECT o.ProductId, c.dates
   FROM [productData.csv] o,
   (SELECT a.StartDate dates
   FROM [productData.csv] a
   UNION
   SELECT b.EndDate
   FROM [productData.csv] b) c
   ORDER BY o.ProductId desc
3. To create Availability table add a custom table with the following script (changing table and column names accordingly):
   
   SELECT DISTINCT a.dates, res1.avail, res2.notavail
   FROM [AllDates] a,
    (SELECT a.dates, count(b.ProductId) avail
     FROM [AllDates] a,
     [productData.csv]b
     WHERE a.ProductId = b.ProductId
     AND a.dates <= b.EndDate
     AND a.dates >= b.StartDate
     GROUP BY a.dates
     ) res1,
    (SELECT a.dates, count(b.ProductId) notavail
     FROM [AllDates] a,
     [productData.csv]b
     WHERE a.ProductId = b.ProductId
     AND (a.dates > b.EndDate
     or a.dates < b.StartDate)
     GROUP BY a.dates
     ) res2
   WHERE a.dates = res1.dates
   AND a.dates = res2.dates
   
4. In Sisense web, add a new Columns Widget and choose in the right pane a "Stacked" display
5. Create calculated measure "SUM([Notavail])" and name "Not Available". Similarly create "SUM([avail]) and name "Available". Pick your colors and you are good to go!